Data Sharing & Data-Driven Decision-Making Task Force
Video
Transcript
5 documents
Machine transcript
May contain errors. Verify important quotations against the official video.
About transcript accuracy
- Source
- Whisper
- Model
- ggml-large-v3-turbo.bin RTX5060
- Processing date
- October 8, 2026
Speaker 1
0:00
So glad to see everybody here. It's kind of like old home week. We haven't seen each other for a while. Glad to have everybody here. We have a few more
Senator Jane English
Unverified
0:09
folks to join us, I hope. And so we're back to our data-driven task force, our transparency task force. And I think one of the things we need to do is make sure we kind of understand what our goal for the task force is. So one of the things I think you can see right over here, I have this thing, and you also have a handout on your desk,
is when we did the Workforce Education Excellence Task Force, one of the things we did was some asset mapping. So what you can see over here from your handout is we've got all these little agencies, all these agencies with all this money, with all these different programs, but nothing is coordinated. We don't have any idea what the results are of any of the dollars that are being spent, they're just being spent. So we as legislators appropriate,
continuously appropriate money, but we have no idea what the results are. So part of transparency and getting people to be able to share the data, is to be able to look for what are the results? How are programs working together or are they not? Are they duplicating services? Are there things out there that they could be doing better? So this is just kind of a little overwhelming if you look at it and think about, we don't know how we're spending almost $5 billion,
and we don't know what the results are of that. So it goes all the way through kindergarten, through careers, federal programs, state programs, all kinds of things like that. So one of the things that we have to do, and from my standpoint, I've said all along, I don't know what I don't know. And probably all of us are kind of in that same boat. But one of the, in the newspaper today, was an article about this healthcare group that's pulling data together, somebody's doing it,
pulling data together to begin to look at the opioid use and where it is and all that. Well, that's the same kind of thing that our Department of Information Systems should be able to do for us, is to do that, get that data and put that stuff together so that we can make better decisions. We don't know what the issues are right now. Don't know exactly what all the business cases are, but certainly we can begin to find that out. So I think from my standpoint, one of the things that is probably the hardest is
we have all these people sitting on this transparency data panel. Everybody has information in their little purview. We've got to figure out what's the mechanism for
being able to share that information across
party lines. Everybody wants to hold on to it. Well, at the end of the day, it isn't a matter of are we going to do it. It's a matter of how are we going to do it. And we have to come up with everybody's help in figuring out what a governance thing would look like,
What would a hub look like so that at the end of the day we can suggest something, put legislation forward? We also need to figure out what the funding is and why it is. So I was telling somebody this week I have spent a lot of time on my little iPad looking at stuff like machine learning. I don't know anything about machine learning. I keep hearing about it. Well, I know enough probably to be dangerous right now.
But I did pick up a lot of things that one of the things that the department has said is that that's an important part. Well, right now, they don't have that capability. What would it take to get that capability? So at the end of the day, we have a lot of consultants doing a lot of gathering of data that we're paying a lot of money to. We should be able to do that inside of state government. So, one of the things we have to do is we have to vote on the last minutes.
Do I hear a motion? Second? All in favor? Aye. Thank you very much. Okay. So, now I'm calling on Ms. Jessica Jones. She's going to make a presentation here. It's really hard to
see. I thought this would be a better room, but it's really not. Why don't we take these things off of the table? Can we take these? We know who we are.
I think. Thanks. Just in case you don't know who you are. Thank you. You didn't identify yourself. You'll be recognized. Thank you,
Jessica Jones
Unverified
5:00
Madam Chair. My name is Jessica Jones, and I'm the Director of the Arkansas Department of Information Systems. And if it's okay, I would like to bring
Speaker 15
5:07
Arata Karkura, who is our Deputy Chief Data Officer, who's going to help me also with the presentation.
Speaker 16
5:19
Thank you so much, members of the task force. It's a pleasure to be here in front of you again. I think last meeting that we had, we had great conversations and there were great questions. But I thought it would be very pertinent for this meeting to talk about the work that has been happening with the data transparency panel as of today. There's a lot of work that has been going on for the last several meetings, and I'm going to go through what we have done so far.
The responsibilities, and you have on your handout, we put together the responsibilities of the data transparency panel. And there are lists there. This data transparency panel was created by Act 912 of 2017. It created the positions of the chief data officer and the chief privacy officer, chief privacy officer, excuse me. Some of the responsibilities was to establish a data catalog. We really needed to see what kind of data is out there, who's the owner of the data,
how's the data regulated, and so that when there's an agency that requests data from another agency that we can see if in fact it can be shared or not, or where is that data. The other one was to do a feasibility and cost study of the development of a statewide data warehouse program, determine and recommend procedures necessary for the implementation of a statewide data warehouse, oversee a statewide data warehouse program implemented in the state,
evaluate and identify data that might be provided to the public when we talk about open data, master data management to facilitate data standardization, deduplication, sharing, data integration of critical data between systems and state agencies, and there's other responsibilities. I don't want to read the entire slide. So, these are some of the responsibilities that the panel has been tasked to, and what I want to do, I want to go through what the work that we've done so far,
because that's last meeting with the questions that were raised is some of the work that we currently have done. On slide three, you'll see the members of the data transparency panel. I believe we have about 13 state agencies. We have the chief privacy officer, it's chaired by the chief data officer. There's a couple of appointees by the governor, by the president pro temp, and the speaker of the house. This last force I'm gonna talk about now,
there was, we did a gap analysis of the state department needs. This gap analysis was completed in collaboration with ULR, the Institute of Chief Data Officers, in collaboration with DIS. So there was a listening tour for executive branch agencies. Those are the agencies that are part of the panel, so those 13s. The CDO, the Deputy CDO, went into what we call a listening tour, sat down with agency from agency heads, the data stewards,
to figure it out what are the data needs and where are the gaps.
Speaker 18
8:36
We documented the identified gaps, and we provided recommendations on action plans. On slide five, we have the gap analysis findings. We have a little chart right
Speaker 16
8:53
there. The analysis identifies six gaps in data needs and management. The gaps identified were assigned a priority based on the number of times the need was identified by an agency as its top priority.
As you can tell data sharing was one of the big needs. We have about 50% of the agencies talking about the need for data sharing followed by data integration. When we talk about data integration is the ability to identify an individual across agencies. This, based on these findings we have recommendations, there was follow-up recommendations on the creation of a multi-agency data sharing agreement, so that was one of the recommendations based on these findings, the implementation
of master data management system and a master citizen record, and the adoption of a data governance model and data quality management standards.
Speaker 18
9:52
So when I talk about the creation of a comprehensive multi-agency
Speaker 16
9:57
data sharing, we're talking about based on the data catalog, now we can see what kind of data can be shared and the privacy officer will have a lead on putting together the multi-agency data sharing agreement.
We talk about the implementation of a master data management system to manage the critical data of the state and to provide data integration in a single point of reference. We talk about the creation of a citizen master record to ensure consistent identification of citizens across agencies. We talked about the adoption of a data governance model to be able to manage the availability, usability, the security and integrity of the data in our enterprise
across the state. And we talked about data quality to ensure that highly-quality data data exists across the enterprise. So those were the findings
Speaker 18
11:01
on the gap analysis, and I'm going to turn it to Adita, who's going to talk about the efforts
Speaker 16
11:07
that went on in creating the state data catalog.
Adita Kerkera
Unverified
11:15
Good afternoon. My name is Adita Kerkera.
I'm serving as the deputy's chief data officer with the Division of Information Systems. Thank you, Director Jones. As Director Jones was talking about the different initiatives that were taken on by the CDO office, one of the first things that we realized was that the state did not have a collection and inventory of the data assets that were hosted and housed with different state agencies. So we can't really leverage data that we don't know exists. So our first step was to go out and start talking to the agencies to help them understand why we need to start creating an inventory of the data that you host.
This was not only a need that was identified by us, but it was also mandated by Act 912, which created the position of the CDO, the CPO, and the DDP. The act specifically mandated that we need to collect the data that is housed with the different state agencies. It was also a recommendation by the gap analysis that Director Jones was just talking about. So several reasons why we needed to go ahead and start doing this effort and collecting the inventory of the state data that is hosted with different agencies and departments.
Right now, as of today, we have been working on this effort for about eight months. We have collected data from the data. And when I say data, we're not collecting actual data, but these are data assets
that are owned by an agency. So not actual data, but the metadata about that data asset. Yes,
Senator Jane English
Unverified
12:45
ma'am. So tell me what that is, data assets. What kind of a form are they in?
Speaker 1
12:50
They are just records of a system, of a data system.
Adita Kerkera
Unverified
12:54
So if, for example, ACES, ACES is a system of record maintained by DFA. Not I mean, yes, I'm using ACES as an example because that is the most common one. But different agencies have different data systems that they are hosting. But not everyone knows what they are. If we don't know what that system is, we don't know how we can start to connect to it and how we can start leveraging it for making useful decisions. So let's say the Department of Human Services may have a data system about Medicaid.
DIS may have information about circuits or the state data center. So different agencies have different data system and all that we were asking to collect with this was an inventory of what each agency was hosting. Not actual data, but just an inventory of what
Speaker 16
13:44
you're hosting. Okay. If I may, real quick, if you can look at slide seven, all we're doing, we're just asking agencies, what kind of data do you host? Where is that data? It's
Jessica Jones
Unverified
13:57
just really an inventory of the data.
So if you look at it, my graphic is pretty
Adita Kerkera
Unverified
14:04
small. Yes. Can you read it? Yes. Sorry. but I think the critical information that I may point out is the agency name, the name of the data system, a brief description of the data system, the data steward for the data system, the business owner, whether the data hosted in that system is sensitive or not, and if that data is publicly available today or not. These are some
Speaker 26
14:30
of the attributes or some of the values that we are collecting regarding
that data system. And I think the main purpose of this was so that we can start to know what different
Adita Kerkera
Unverified
14:41
agencies have today. Somebody needs to know in this state what all do we have sitting in different departments. If you don't know what we have, we can't start connecting the dots to start leveraging them for better decision making. So we have pretty much all the agencies that are part of the panel. They've already started using the system. It's in an online portal today. They're logging in and maintaining the data systems that they're hosting, making changes
if there are changes to their data stewards or changes to the regulations that are surrounding that data. So if a data set is, you know, has to be compliant with FERPA or CJIS or HIPAA, are there any regulations surrounding it? We want to know about it upfront so that when you start talking about sharing that data, you know
Speaker 26
15:27
how to make that data sharing compliant and make sure that the data sharing is secure so that everybody feels safe with
Adita Kerkera
Unverified
15:36
data sharing. So we have already been using this, like I said,
for the last seven, eight months. And most recently when we did the recidivism POC and we're now working on the workforce analytics POC, we've been using this data system to try to see what is hosted in which agency. Right now, the data asset is available only viewable for each agency to view their own data. the CDO office is able to see all the data assets and provide advice to others as
Senator Jane English
Unverified
16:08
needed. The question here, Representative Love has a question.
What I'm wondering is, so if you came up with a business case, let's just use that opioid thing, would you be able, from what you've collected here or you know what an agency is collecting, be able to say these are the 10 different places, the coroner's office, the health department, that would have these kinds of assets that would need to be part of a look-see? That would
Jessica Jones
Unverified
16:37
be the vision for it, and it would be driven by the business owners.
Senator Jane English
Unverified
16:46
Did that answer your question? No, not really. Apologize. I mean, somebody I think would have to come up with, this is the kind of information we need, And then, I don't know, it's a hub or whether it would be you all or whoever it is, would contact those different agencies then, and we're assuming they have that information like coroner's reports or stuff from the health department. But would you have a list enough to know where you have to go
and who might have based on what you've collected here? That is
Speaker 16
17:19
correct. That is the vision to have. Like I said, this is just a listing. When we talk about inventory, we're just talking about a listing. But not specific. Of data assets. we don't have access to the data, we're not accessing, it's just a listing. They can just tell us, we have this data asset, which stores this kind of information for this system. It's just a listing. So yes, the future will be, if there's a business case, that the agencies are able to access this listing or inventory to say, okay, for a particular business case, I will need
data that contains, I'm just going to name an address. I mean, this is just very simple, but I'm just throwing it out there. So then they can go back and look what data assets contain those attributes and who is the owner and how is that data regulated. Can it be shared or cannot be shared? So Representative Love, you had a question? No,
Speaker 42
18:05
I'm missing it. Oh, okay. All right. Let's see.
Representative Austin McCollum
Unverified
18:13
Go ahead. So why wouldn't you want the agencies to be able to view all the data in its entirety?
Why would you just keep it? in one
Adita Kerkera
Unverified
18:25
centralized location? We want the agencies to be able to view all the other agency's data. But I think this was our very first step towards consolidating, not consolidating, I'm sorry, but getting the listing of inventory together. So I think there was some hesitation from the agencies to just open up and present what they have to everyone. But I agree, that should be our next step. We should be able to openly display to every agency what each other hosts.
Speaker 26
18:52
That's the best way to leverage the data. And that is something that we hope
Senator Jane English
Unverified
18:58
to propose in forthcoming meetings. So let me ask you this question, then, and I'll get to the other of you.
So let's say an agency is using a consulting company to gather all this data and do whatever they do with it. Would that be something that would show up on your list? So just like Medicaid, if there's somebody that's gathering a whole lot of information
on Medicaid, would that stay in there with that consulting company or whoever's doing it for them, or would that be part of the asset
Adita Kerkera
Unverified
19:33
stuff? So when we started the effort and we had several sessions with the agencies to kind of identify what are high-value data assets for you, we left that choice onto the agency to say whether you are hosting that data or you're utilizing that data. The key point was, is that data critical for the functioning of your agency? And if it is critical to you,
then it's critical to the state. So now, if they were utilizing that data, but it's stored somewhere else, it was a choice for that agency whether they enlisted it
Representative Fredrick J. Love
Unverified
20:10
in their inventory or not. Representative Blow. Thank you, Madam Chair. Director Jones, now this kind of piqued my
interest because when you started talking about data sharing agreements, why would we want other agencies to see the data from like say for instance
why would we want ADH to see the data from DHS, especially if it's sensitive data? You know, I mean I followed you when you said that you wanted to know if it was under HIPAA or whatever. But when you begin to say, well, that system is opened up to another agency, that could lead to some issues. I mean, I can understand having an agreement to where you all can pull the data and do analytics on it if there's a business reason.
But to have a system just open, I mean, or clarified if I'm getting something wrong, because I thought I heard you all say you want to have less than open data sharing. So, okay. No. You mentioned just a second
Speaker 15
21:18
ago, you pointed out it's going to be business case by business case. And we're not talking about open data for everybody. There's cases for open data that it's open data to the public.
but we're not talking about that. We were just talking about business case by business case. What we're talking about here is just, again, it's just a
Speaker 16
21:37
listing. It's just so that we know who stores what, who's the owner, how's the data regulated. It's just a listing. That's all there is. For right now, and like she mentioned, this is just the first step. This is something else that is part of the
Speaker 18
21:51
recommendation and part of the best practices and also what other states have also done. Okay.
Representative Austin McCollum
Unverified
21:59
Thank you. Yeah, I do realize we're just talking about first
steps right now, but I do want to make sure I'm understanding part of this dialogue. Why would we want it to be the case that you have to have a business scenario to access the data? Why wouldn't you want an open data warehouse to the state agencies? And also, who would decide what a business case scenario is? Who would weigh that?
Speaker 16
22:29
I think that will be based on the agencies. I mean, they will have to have a conversation director to director. To me, when we're talking about business case by business case, we're really talking about analytics, and that's business case driven. We're not talking about having a system where everybody can just go and access. When you talk about a business case, it's collaborating with those agencies and connect what we talk about connecting the dots and connecting the data assets or the data sets together for a particular purpose.
Jessica Jones
Unverified
23:03
They have to have a particular purpose as to why they want to access the data. There has to be a business case behind
Senator Kim Hammer
Unverified
23:16
it. Thank you, Madam Chair. Okay, so not to be redundant, you're just trying to get a catalog listing of all the services available as a first step, correct? Okay. And we don't do that already because it seems that kind of catches me off guard. That seems so common sense that we don't already have that sitting out there somewhere.
Adita Kerkera
Unverified
23:34
We would have thought the same, but really when
we started talking to different agencies, not all the agencies were collecting that data. There was not an inventory of what data they have. So we said that is our first step and, you know, that's what's needed. We can't leverage if you don't know what we have. So we were surprised too, and that's why this effort was started. And today there is a listing available of the agencies in a portal, but it is a
Speaker 26
23:59
closed portal just now. It's not open to the public yet.
Senator Kim Hammer
Unverified
24:04
All right, and so once you get through this hurdle, as you move this project along, or as the committee moves this project along, and it gets down to where there's sensitive data that would be shared in order, for example, us to make good decisions as legislators. We want that good solid data collecting from all. There'll be a gatekeeper, for lack of a better term, that's going to control access to that data that may be sensitive. Is that your kind of ultimate goal that you would push toward?
Adita Kerkera
Unverified
24:34
Yes, but again, we're not talking about any data here just now. We're just talking about an inventory of the data system, not the data. But yes, today the CDO office plays a role of that broker or the gatekeeper that we're talking about. So the CDO office has access to all the inventory of different agencies today, because somebody has to be the gatekeeper. But at some point, and moving forward with this task force or with the data and transparency panel, we hope that we're able to let all the agencies take a look at the data system stored with other agencies.
Speaker 26
25:05
That's the only way they can start sharing and
Senator Will Bond
Unverified
25:16
connecting the dots. All right, thank you, Madam Chair. Thank you. So in this pulling everything together, I have some minor understanding of the Arkansas Research Center, who's been out there doing, I guess is our longitudinal study group for a lot of education data. Are we pulling them into these discussions?
Because they've built out some of these things already. They have six groups they're working with. Are we coordinating with them at all? Are they
Speaker 16
25:50
sitting on the sidelines or what? Apologies. It's probably been about a year and a half ago that we talked to the research center. I can't remember their name. But no, since we're just talking now for this kind of conversations, we're just talking now about the research and the work on the data catalog. No, we do have an insight on the data that they utilize for their efforts.
Senator Will Bond
Unverified
26:14
right so my understanding they have five data sharing agreements I never know whether to say data or data so I mean if somebody is there either way all right that's Jane or Senator English might not have put me on the committee knowing that I couldn't pronounce the word but so they have data sharing arrangements my understanding with five and ADE is not sharing with them right now, is ADE's not sharing with anybody, Arkansas Department of Education?
Jessica Jones
Unverified
26:46
No, not, well, I cannot answer for them. For that particular effort, you're absolutely correct. No, they're not sharing the data with
Senator Will Bond
Unverified
26:52
them. Okay, and they haven't shared since 2013, and I think that's something we have this group out that already has doing longitudinal studies that started out under a federal grant, and that we need to somehow involve them in the discussion of this process already there's already some models they've built under a different system than we're looking at i think but i think that expertise needs to be utilized and figuring out what's out there and
then figuring out how we're going to analyze it uh and i agree with senator hammer this idea that we didn't have a list of where all the data is and we're not utilizing all our assets or getting everybody on the same page is something that i think hopefully we're going
Jessica Jones
Unverified
27:36
to move move to do. Right? So, hopefully. MS. Definitely. I mean, definitely for sure. I
Speaker 16
27:41
think, you know, this goes beyond reporting and beyond that. I mean, this goes to we are pursuing not only the analysis of the data.
Jessica Jones
Unverified
27:48
We're not just talking about aggregating and we're not talking about reporting. We're talking about utilizing predictive analytics. We're talking about utilizing machine learning to leverage and to really come up with insights, again, based on business cases. Right. Representative Brown. I'm sorry. No, no.
Senator Will Bond
Unverified
28:06
Thank you, Senator. I may have some more later.
Senator Jane English
Unverified
28:09
But I guess what you can see is we don't know what we don't know. So this is really a real learning thing for all of us to figure out what this all is. Representative Brown.
Speaker 80
28:19
Thank you, Madam Chair. You made the comment a moment ago that we are only going to be accessing the data that's housed within the agencies or the types of data that they have, but that we will not be accessing what their consultants may have. Is it not important to know the data, the types of data that the consultants are gathering
so that if we decide we need that, we can get
Adita Kerkera
Unverified
28:56
it? It's just a question. I think there is value in collecting high-value data assets that may be collected by third parties for an agency. And that's why we gave that choice to the agencies to say, this is the first time we're trying to collect an inventory. We gave them the choice to say, you need to come up with your high-value data assets.
What are highly critical data assets that we should be keeping an inventory of? And in some cases, they put down third parties, and in some cases, they did not. But we do not have any hard and fast rules written down to say, yes, you have to log every data that is being collected by
Senator Jane English
Unverified
29:40
a third party. Do we know how many agencies are having data collected by
Speaker 82
29:45
third parties? We don't know that information. And
Speaker 16
29:48
I would assume that, and a lot of agencies have other vendor partners that do great work for particular agencies.
So, again, I'm just generalizing. I'm not saying that's the norm. But I would assume that those vendor partners are utilizing the state data to do whatever business process that they need to do. So at that point, we do have the information that is currently at the state. What the vendor partners do, whether it's reporting, whether whatever it is, the main core or the input to them, we
Senator Jane English
Unverified
30:25
have that information. Thank you. So if it's Medicaid, we have that Medicaid information, they're using it?
Jessica Jones
Unverified
30:33
Would that be true? I cannot say for sure. I think, you know, and that's a great point. I think we
Speaker 16
30:40
need to bring that to the data transparency panel to bring that in front of the state agency so that I cannot answer for DHS. Fine. I can only assume. But, yes, I mean, we can bring that to the data transparency panel so that it's clear and we
Senator Jane English
Unverified
30:56
can bring that answer to here later. Because if you really wanted to use that data, if you were going to look at tracking somebody from kindergarten through death,
you would need to have some of that Medicaid information. Or if you were thinking about what happens to people, how do they enter the prison system, or where do they go from place? You almost need all of that information. What were those places, those on and off places that those people, you almost would need that information to be able to make those kinds of decisions, as long as we've got the information and they're just using it to do whatever they do with it.
But we don't really, I guess the agencies know, but
Jessica Jones
Unverified
31:43
we don't. And Madam Chair, I'll take
Speaker 16
31:45
that as an action item to address it in our next data transparency panel so that we
Jessica Jones
Unverified
31:51
can bring that question to the rest of
Senator Jane English
Unverified
31:53
the agencies. Because what they're doing, we may be able to use some of that for ourselves for some decision-making. Who else is? Senator Hammer, did you have another? Thank you,
Senator Kim Hammer
Unverified
32:07
Madam Chair. I'm just curious. I see Roger in the room.
So are you incorporating into the discussion legislative audit and OMIG to make sure the compliance, that there's good communication between what we're doing so that if Roger and legislative audit has to come in and audit something, that we're all on the same page going forward? No, that I'm not aware of
Representative Fredrick J. Love
Unverified
32:40
Thank you. Representative Love. Thank you, Madam Chair. Director, you all keep referring to the term
gatekeeper. And it's my understanding that you all are not going to be collecting data.
So tell me what goes into that term when you're saying gatekeeper?
Jessica Jones
Unverified
32:58
I think what Ms. Karkura referred to when she mentioned the date keepers, she was talking about the inventory. I'm going back to, again, I'm going to try to bring it back. We were talking about the data catalog. And when she mentioned that, as we just started in the process, and like
Speaker 16
33:17
Representative McCollum mentioned, the inventory, or I'm going to call it the listing. I mean, I'm just, to simplify, just think about a spreadsheet that contains the listing of who owns what and who's the owner. Today, because we just started the process, only the city office has access to the entire listing.
The agencies can see, per the agency, the data asset, the owner. The vision is to open it up for the agencies to look at that listing, to figure it out who owns what. Again, we will bring that forward to the data transparency panel to make sure that that's approved, that we can open up that listing, or I call it, it's not a spreadsheet because it's a portal, but I'm just going to call it a spreadsheet that everybody can look at. Okay. All right. Thank you. You're welcome. Senator Sturge.
Senator James Sturch
Unverified
34:08
Thank you, Madam Chair. And, Director, I may be transitioning to your next slide, and if I am, I apologize. But
one of the gaps that you mentioned on your spreadsheet, and we talked about it last meeting, I think it was the Chief Privacy Officer who talked about the comprehensive data sharing agreement between multi-agencies. And basically, no offense to my friend Senator Bond, but she made it sound like getting lawyers or counsel from different agencies to agree on something was proving to be very difficult. So where are we in the master agreement where all agencies would agree to a data sharing
agreement that everybody would be comfortable with? Have we got it drafted? Is it going to have to be legislatively mandated? Where are we on that, and is that
part of the discussion from activities from other states when we talk about Indiana, for example?
Speaker 21
34:55
Thank you. And this is, and I'm going to jump into kind of one of
Speaker 16
35:01
my closing items, because this is something when you talk about the DTP, the data transparency panel, the main reason why I listed the responsibilities is because you can see what we've done so far.
So we can do certain tasks. We've been tasked to do the data catalog, the cost feasibility study, and several things. However, we have no authority nor funding to move forward with the recommendations of the gap analysis, with the recommendations of the cost feasibility study. There's the reason why this task force exists today. So, and Senator, if I may, you know, based on the recommendations and once we can provide those to you, that will be up to this task force to decide how do you guys want to move forward with this effort?
Is it going to take legislation to put together data sharing across agencies? Is it going to take legislation to put funding for this effort? Senator Bond, do you have another? Well,
Senator Will Bond
Unverified
36:01
and Senator Sturge makes a good point. That's what I was getting around to, too. It seems to me that the data sharing agreement, and I know there are technical legal aspects to it, but that there's some sense with some of these agencies, if you hold on to your data, no one else can analyze it. But, you know, it sort of creates a silo effect, and I think we ought to be pushing hard that that gets out.
I'm shocked that ADE isn't sharing their data right now because they're grading their own paper to some extent. You know, I mean, it doesn't, and I know FERPA and all these things exist out there, but look, we all waive different things all the time. I mean, there's ways to deal with these issues. so I don't want to I think that's incredibly important and something that we need to address and force this data sharing agreement on everyone. I think that's probably the crux of what it
Senator Jane English
Unverified
36:49
is we're going to be trying to do we just have to figure out what we're doing but I think that's
exactly what it is is how do we come up with something that gives everybody the comfort level and as James said sometimes you get lawyers sitting in a room and tell you 40 different reasons why you can't do it right so for
Speaker 97
37:08
me particularly when they don't want to do it you know that's my experience so
Senator Jane English
Unverified
37:12
for me this thing is not are we going to it's how do we do it and so that by the end of the year we have we're firm in our own minds what it is we're doing and what the options are whether it's an executive order whether it's legislation or whatever
and and the funding and we need to have a better understanding of what does that look like what does that funding look like? Well, and that's something I
Senator Will Bond
Unverified
37:35
was going to say. I know other states have tried it, and maybe I don't know if we have the resources to bring
some of those people in to say, okay, how much did it cost you? What were the pitfalls? And what did you do? Yeah. Thank you. Yes.
Senator Breanne Davis
Unverified
37:50
Senator Davis. It's my understanding that you guys maybe
have done that, brought in chief data officers from other states, and to sort of talk to you guys as you went through this process
over the last year and a half. That's part of our
Senator Jane English
Unverified
38:04
presentation. So that may be something that we need to hear, though, that same kind of discussion, though, as the members of the task force may need to hear that, because we're all operating in a different world than you all are, obviously, because that's not our area of expertise. So it probably would be very helpful for us to hear from other states, because I know, I see it all the time, that they're doing things, so we might. And particularly on costs,
Senator Will Bond
Unverified
38:27
I think it might be helpful to hear from what did it cost these people, what's the real cost of it, then again?
Senator Jane English
Unverified
38:34
Well, and the other thing is we know it's going to
cost some money, but the reality is then is the cost offsetting, I mean, is that small in comparison to what we might be saving? I mean, to me, this is transformative. If you can start getting agencies to talk to each other and use data to make much better decisions, one of the things that you had mentioned before when you were talking about the recidivism was something that we looked at or we talked about was how the Department of Corrections
was doing some stuff with mental health, and DHS was doing stuff with mental health. Well, if they'd have looked at what both of them were doing, they might have been able to do a better job of figuring out where to locate those mental health things so that they could be beneficial to a lot more people. I think that's kind of what we're looking at. It's not just about dollars. It's about how do you use the resources better. So I'm sorry we've interrupted you.
Adita Kerkera
Unverified
39:38
I said we don't know what we don't know. So that gets us perfectly on to our next topic, which is activities in other states like we were talking about. So I'm going to jump around the agenda that I had in mind just a little bit in view of the question that was posed earlier. Probably the most widely recognized and successful example of transforming state government with data is Indiana. So we talked about Indiana just a little bit last time. Their management performance hub is probably the best example of getting data together.
They are functioning with an executive order for data sharing. So, you know, that's an example that we can look at. It's the most successful model with an executive order for data sharing and a functional center of excellence for analytics. They worked on studies to guide policy and increase transparency for several use cases. Some of them are similar to the issues that we're facing with recidivism, income tax fraud, highway crash prediction, Medicaid optimization education and workforce, and the opioid epidemic.
Indiana, statewide longitudinal data system is also hosted with the same management performance hub. So that's definitely one of the models that we've been closely following. and it is in place with an executive order for data sharing. Last, in November 2017, the CDO office convened a CDO forum. We call it the inaugural CDO forum. And that was our first step to start discussions on interagency data sharing
and also to start to highlight why data is a strategic asset for data-driven government. It was very well attended with all agency leadership from within the state. Governor Hutchinson, in fact, gave the opening keynote for it and talked about how data should be used more and more as a strategic asset for decision-making to be a more successful data-driven government. We had three CDOs in-state for that forum. We had the CDOs from New York, Texas, and Tennessee come to us
and talk to us about not only the challenges but also the successful stories that they have faced back in their states. So it was a great learning opportunity for our leadership to hear what's worked and what's not worked, how their funding mechanisms were, what kind of resources did they have available, both personal and technology-wise, to get some of the deliverables that they were able to achieve. So I'm not going to go through all of them, but some of the highlights of those discussions were Texas has a statewide data exchange compact, which is just a big word for a statewide data sharing agreement.
So they have a standardized set of terms that are used in an agreement with them, which they use for data sharing between different agencies. So we've been working with them to see how that's working out for them. Most recently, Texas has been focused on leveraging a third-party solution for an open data portal. And they're using that not only to publish some of these data sets that we were earlier talking about that are not sensitive, so they are openly publishing those out and making them available to the public. The same work has been done by New York.
They have a very successful open data portal, so there are a lot of data assets available, not only the inventory, but actual data available. Any data sets that were not sensitive are openly available in New York's data portal. spoke about a statewide master data management solution that they had deployed and they have a really interesting model they are they were proposing using a centralized pool of resources they have a centralized pool of data scientists that they use in a shared service model to work on problems
that are faced with different agencies so for cross-agency work and for agencies that can staff specialized resources they have a pool of data scientists and data architects that can together work in a central agency and still provide services to other agencies that can't afford those resources. So interesting model there. Other than that, we've also been in touch with Arizona. They are also working on an enterprise data sharing agreement. Florida, we worked closely with them while we did our catalog of our data assets.
In fact, Florida and Arkansas pretty much completed their inventory at the same time. and we worked very closely in our processes and our methodology of collecting the data assets. Connecticut has an excellent open data portal as well, and they are currently working on a state data plan. And they're publishing not only the data inventory they have, but they're also starting to publish their open data as well. Maryland has deployed a similar open data platform as well.
So just kind of giving you an idea of what's happening in the other states, we are constantly in touch with them, with the leadership in the data area for different states to kind of learn and exchange what's working in their state, what's working in ours, and continue to learn and grow with each
Senator Breanne Davis
Unverified
44:49
other. Have you guys been speaking with North Carolina at all? I know they have a government data analytics center, and they have centralized all their data and had business use cases for
Adita Kerkera
Unverified
44:58
it. I don't believe we've been in touch with North
Representative Austin McCollum
Unverified
45:01
Carolina but I will add that to my list
I'm curious to hear out of what you've learned from these other states if if you think one is doing it better whether it's using like in-house data scientists or some of those third-party agreements that they have other pros cons either way of after speaking with these other other states that you could kind of guide our
Speaker 26
45:31
thinking to, or? Well, we're seeing a mix.
I know, like I said, Tennessee followed a shared resource pool
Adita Kerkera
Unverified
45:36
methodology. So they have a certain pool, and they are able to use these resources for cross-agency use cases that come up. And I think that's worked really well for them. We have not been in touch with Tennessee very recently because the last CDO who was here in state for that visit is no longer there and they don't have the position filled. But I would be interested to see how that's still working out for them because that's a really interesting concept for a shared service model.
It's been different things that's worked and not worked in different states, but one thing seems to be common, the issue of data sharing. the issue of needing a centralized master citizen record, because every state is facing the same challenge of inconsistency and inaccuracy in the citizen record, whether it's name, address, and so forth. That's one common problem that all the state CDOs talk about every time we get on our monthly call to discuss our issues.
Different methodologies are working in different places, different priorities in different states. is definitely an issue also in most states. But they're all working towards trying to get to that master data management,
Speaker 26
46:48
which is really a foundational step for any of these solutions. So, and I keep hearing Indiana and
Senator Will Bond
Unverified
46:57
then North Carolina, that kind of goes back to maybe getting some of those people there. And I don't want to get too far in the weeds, but Indiana has a federated data system. Is that separate and apart from a centralized data system?
I think there's two different types of systems. Do we know what type
Adita Kerkera
Unverified
47:16
they're using if we're going to do it? I believe they're using a federated system, but I'm looking at
Representative Andy Davis
Unverified
47:25
Mr. Magoo, and he might be able to give me a little bit more input on
Speaker 111
47:29
that. They control it all together, identify for the scientists, separate systems
Adita Kerkera
Unverified
47:34
have to be identified. Can you repeat that? Yes. So they have their sensitive data is restricted in a manner where only specific agencies can use it, but the rest of it is centralized.
I'm sorry, I cannot hear you, and I would defer to Mr.
Senator Will Bond
Unverified
47:53
Magood to talk on that. And that may be too down the weeds for this meeting, but I know that's on down the road, but
Speaker 117
48:02
that's why I think I'm interested to hear from some of these people from other states. Yeah. Their architecture, which,
Speaker 79
48:10
and it's technology. You need to identify yourself. This is Robert Magoo, Deputy CDO with DIS. Indiana's approach currently today,
Speaker 121
48:16
which was based out of 2013, and technology of those practices change over time,
so it doesn't necessarily mean it's the best model. They've got two systems. They've got them actually. It's below the governor's office in the capital at a very secure area. They bring all the data together onto one system. That's identified. It's very, very limited access for integrating the data and doing the data science. Once it's been integrated, they move it to a completely separate system where it's de-identified and aggregated to meet compliance needs for open data access.
It's not that we know of unless we talked to them currently federated, but I don't think that's really an intentional decision. I think that was, like, the case in 2013 when they first set it up, just the easiest
Speaker 82
49:11
architectural approach. Thank you. One of the last items I want to
Speaker 16
49:14
cover at a very, very high level, the work that has been done with the feasibility and cost study of the development of a statewide data warehouse program.
This feasibility and cost study was required by Act 912 of 2017. The study was completed by the CDO office in collaboration with some of the members of the data transparency panel. It is important to know that it's still under review and it's not been approved yet. So we will be seeking approval for this study on the next data transparency panel meeting. Once we get approval, we will be happy to share the study with all the members here. The study consisted, of course, on really a research of the current data management and analytics best practices.
It also, the study of other public sector entities, just like Adita and Robert were talking about, of a study of what other states have implemented, pretty much. And the recommendation was to have a centralized trusted broker, what we call
Jessica Jones
Unverified
50:21
the data hub, within the Division of the Department of Information Systems as an agent of individual agencies
Speaker 16
50:29
in integrating and providing compliant access to
an analysis of interagency data. As you
Jessica Jones
Unverified
50:35
can see, this is on slide nine. And again, once that
Speaker 82
50:50
study has been approved, we'll be happy to share with everybody. But
Speaker 16
50:55
this study pretty much will have the answers to your questions as far as what other states have done. What was the cost in different states?
We have different examples, and we'll be able to provide that to you. Again, really, the two items that were here and the current barriers that we're facing, again is the data sharing for consistent cross-agency data reference. And another barrier is a sustainable funding model for centralized data sharing and data driven decision support services. The city office has completed the research and the data and transparency panel accepted the findings for the data gap
analysis, but again under the current law the panel does not have the ability to to fund the implementation of the recommendations of the gap analysis. The task force, of course, has the authority to implement the recommendations of the study and the gap analysis by requiring participation in enterprise data sharing and providing the funding to implement a solution
Jessica Jones
Unverified
52:00
which removes the barriers currently facing agencies and moves Arkansas towards a data-driven decision-making.
Senator Breanne Davis
Unverified
52:09
any questions. Thank you. I'm just curious. You've mentioned the state of Indiana. Do they do this in-house or do they have a vendor that they work with? Do you know? In-house? And I have who is highly
involved in that. Did you say Indiana? Yes. Oh, Indiana.
Speaker 121
52:30
Indiana, yes. It is a mixture. They have a staff of 30, about eight of those are contractors. They They tend to be on the data science side. The rest of the staff, they go around to agencies, do education, really just trying to kind of
raise the analytical maturity across the state and helping to find performance measures for all the agencies.
Senator Breanne Davis
Unverified
52:48
Okay. Do you know what vendor they work with by
Senator Jane English
Unverified
52:54
chance? It's a mixture. A mixture of vendors. Thank you. It's not one vendor, it's a number of people who with this
Speaker 121
53:03
expertise that are hired? With data science, yeah, it's a mixture that's state resources and some contractors. When they first started out, they did not have in-house expertise in data science and machine learning,
so they leveraged some contract resources for that. And some of those, there are fewer today, but some of
Senator Jane English
Unverified
53:21
them are still with them. So tell me what goes into,
I know if you have to develop machine learning, what goes into that? I've seen that there are some small, I mean, quick ways to do machine learning, but then some very complicated ways. How would you use machine learning for some of the business cases that we're looking at?
You all would develop the algorithms and all that for whatever that is that we're trying to figure out?
Speaker 121
53:55
Yeah, yes, ma'am. I think that's one approach. I guess because we saw recidivism, that might be the easiest one to explain it with. I guess the current state is to sort of look at historical recidivism data. And even now, with current techniques, sometimes we have to wait for a study. It might be a year to three years before we know what recidivism was historically.
But we've got 20 years of data as a state. So the machine learning is really learning the patterns and the rules from the data. So that as we get new data coming
Speaker 131
54:27
in, instead of just looking backwards, we can predict forwards and kind of understand what are the actionable levers to change
Senator Jane English
Unverified
54:35
that outcome in the future. Right. So does the information come in, I mean, is it all in the same format? Or does everybody put stuff, put information
Speaker 121
54:46
in in a different way? It varies. One of the reasons we chose recidivism, that particular system was very
relatively clean in the format, very well structured. It was a lot easier to work with. It varies from system to system and agency to agency. It definitely takes longer if it's not in the same format. And that's some of the reasons for the master data management. That's really just a common language. If everyone's representing a school a different way, it's hard to bring that together. It takes extra work and extra time. So it's establishing those kind of common codes and language, so it's faster and easier to bring that data
Senator Jane English
Unverified
55:22
together. MS. So we would encourage all agencies to be using the same common format?
Speaker 121
55:26
MR. Yes, and that's data standardization in a noninvasive way, and that means not causing disruption of having them change systems. There's a way to cross-reference these things and make it available so that we can still talk the same language without, you know, changing the systems that have been
Representative Austin McCollum
Unverified
55:48
around for decades. So a follow-up on that, so you're
saying the better way would be to standardize and smooth the data after the agencies collect it, so do that within the warehouse?
Or would you recommend the agencies do that? And I'm not when I get too technical,
Speaker 135
56:03
because I have to make the stuff work at scale in the real world. And the way we do that usually is through attrition. So if there's systems that are
Speaker 121
56:12
already there, that already have these codes, a lot of times When we talk to the federal government, they specify the exact codes we have to use. So if the systems already exist, they're already using codes, we don't want to disrupt that. We'll map that to the codes that need to
Speaker 135
56:27
be used. But going forward, as new systems are developed or systems are changed,
that's the opportunity to start using those new codes. Do you know based
Representative Austin McCollum
Unverified
56:36
off of that, the inventory that we're starting to do of the systems, how much are standardized already? Not
Speaker 135
56:43
yet. And the way you do these inventories is kind of at the first step, it's just sort of what
Speaker 121
56:49
are the systems? ASIS, very large system, because that's procurement, that's finance, that's HR. So it's, there's a lot there. It has 10,000 tables, but at the first level, you're just, what are your systems?
Because even then, you've got very many, and then you kind of decompose that. We didn't want to start with asking every field, because that's just too much work. It was an effort to get the systems, and then they're not all really worth drilling into. So that way, like Jessica said, on a business case driven, as we need, we drill down to what are called the fields? The entities. What do you have? You know, have a person, have an asset, and then maybe we'll need to get into what are the attributes or fields on that asset. It has a serial number.
So it's, again, we don't want to cause additional non-value-added work and cost. We
Speaker 135
57:37
just want to build a framework to drill down on an as-needed basis. It
Senator Jane English
Unverified
57:45
makes sense. Representative Brown, did you have a question? Yes,
Speaker 80
57:52
Madam Chair. When you say these systems have been in place for decades, when you use the word system, are you talking about the types of data that are collected
or are you talking about mainframes? and really
Speaker 121
58:11
should maybe say more data sets, but these tend to be any type of system. They could be mainframe today. They're usually more client server. They might be a cloud system. So the technology's not really specific. A lot of times there's different levels of control. And if they're using a third party, like a commercial off-the-shelf system, we might not I have the ability to mandate that a certain type of code is needed.
So we don't want to cause the extra cost or disruption of just pushing those codes down. So that's where, with that sort of non-invasive way with the master data management, we can kind of mitigate that by mapping between the codes, by creating
Speaker 80
59:00
that cross-reference or, like, an index. Follow-up. Go ahead. So that's programmer work, right? Programmers do those things to cross-reference and bring, like if I put something in in this system in one way
Speaker 143
59:12
and I want to observe it on this other system that reads things differently, you've got somebody
Speaker 121
59:19
in the middle who can It's actually more like business subject matter expert work. That's where we're talking about the stewards because you have to know the business, the real world entity you're describing, the data. So programmers might, you know, facilitate and support the application to allow that interface for the business user, but it really comes down to if someone's mapping something about a person,
it might be someone in human resources or if it's,
Speaker 135
59:45
you know, a financial structure, it's someone in accounting.
Speaker 80
59:52
Thank you. And just one little question. And are mainframes still, are we still actively using mainframes a lot? Ms. We have. Thank you. Yes,
Jessica Jones
Unverified
1:00:05
we have two agencies that are currently using mainframe. There's a
couple of applications that are hosted in the mainframe, but the agencies have a roadmap to get off the mainframe in the next, I believe, three to four years.
Speaker 91
1:00:17
Ms. Thank you. More questions here? James? Senator Starch. Thank
Senator James Sturch
Unverified
1:00:24
you, Madam Chair. Something I can't remember which, if it was Director Jones or Deputy Director Carver, but you said you didn't have the resources or the authority to do anything as far as funding goes, but did you go ahead and estimate what the cost would be or what the, because on your last, on this last slide about feasibility, it says cost and benefit estimation. So was that completed or is that something you're still liking to
Speaker 16
1:00:48
do because of lack of resources? The study is completed. The reason why I didn't put it here is because, again, it has not been approved by the Data Transparency Panel. Oh, I see. So I really want to have the input. The members of the Data Transparency Panel, the majority of the members are designees from the agency directors. So typically is their CIOs or their IT personnel. So we always seek their input on that. So they have the study we presented to them at the last meeting. But, of course, you know, now we're going
Jessica Jones
Unverified
1:01:16
to go in and seek approval. or if there's any changes.
So that's why we don't want to put it here until it's
Senator James Sturch
Unverified
1:01:23
finally approved. But when it is approved, then there will be, as part of that report, an estimation of the cost for setting up a data hub as well as what else? That
is absolutely correct, yes. And it will be specific, like $4 million or
Senator Jane English
Unverified
1:01:36
something? Whatever that is. But what does that $4 million look like? Because I think it's pretty hard. If you just throw a number out there and you don't have anything, any plan or justification for it, Would it be helpful for us to meet with that transparency panel?
Senator James Sturch
Unverified
1:01:51
Would it help anybody in here? I'd be glad to get the report when it is finished, but that's
Senator Jane English
Unverified
1:02:01
up to the committee. Okay. Any other questions? Okay. So do you think that report will be ready
Speaker 16
1:02:09
by next month? Yes. We should be. The next meeting is July 29th, and again, that's where we're going to be seeking approval for
Senator Jane English
Unverified
1:02:17
that report. Okay, so do you all want to have somebody come from any of, like, Indiana to kind of talk to us about?
I mean, it's not that we don't trust what they're saying. It's just kind of like a different voice. Is that something we'll? North Carolina or Indiana? North Carolina or Indiana? Both. Both? We can do that. That might be what
Senator James Sturch
Unverified
1:02:41
we would try our next meeting. Our privacy director, I think it was. I'm sorry. I can't remember her name. But the lady who's over, kind of our privacy director, she was the one, I think, that first mentioned Indiana. And she actually knew some of the specifics as far as how much money they had put in, how much money was saved the first year.
So I think it would be great if we could get, you know, someone from Indiana to kind of show us, you know, because now that's been over two years ago. So it would be nice to see how much they've done. And
then, like you said, not how much, but what did they
Senator Jane English
Unverified
1:03:16
get for the money, you know, and what was that? Well, you know, they started off, and I don't know if it was Pence when he was there or it was before Pence, but I know that they really started doing a real look at state
government and where the dollars were going and how many people we had working here. I mean, just really looking at the programs to determine, and I think a lot of that, of what we're talking about here has come out of what they were doing there because they really did start going through state agencies to determine is there are we have three people at three organizations doing exactly the same thing and maybe with a different funding source
but they were all doing the same thing and how and so some of this transparency and i think Jennifer sent me a thing, too, on their, was it Texas, maybe, transparency site. You know, we've got our own transparency site, but this kind of goes even further than, and that's another thing. We haven't looked at that one for five years. Five, six years it's been in place, so that might be something else. Well, we'll try and see if we can get somebody from Indiana and maybe North Carolina to come
and visit with us. Maybe if the report is finished the next time, that will be on our agenda, so we can talk about that.
Senator Breanne Davis
Unverified
1:04:45
If you don't mind, Madam Chair, that it may be good to hear from two different, just because every state starts with a different business use case and a totally different agency. Why did you do this? It would just be interesting to hear some different viewpoints on why they started where they did. Senator Johnson, did you have a question? Oh, okay.
Senator Jane English
Unverified
1:05:09
All right. Anybody have anything else? Okay. Thank you
very much. We appreciate you as always. As we say, we're all learning. We don't come from the same world you do. And so we're all learning here, trying to incorporate it into our brains. Thank you very
Speaker 95
1:05:28
much. Oh, gosh. We're finished. Thank you.
Agenda
Call to Order
Comments by the Co-Chairs: Senator English and Representative Meeks
Consideration to Approve June 28, 2019, Meeting Minutes [EXHIBIT C]
Review of Task Force Goals and Responsibilities - Senator English and Representative Meeks
Presentation by the Department of Information Services - Ms. Yessica Jones, Director, DIS
1. Overview of Data and Transparency Panel (DTP) a. Membership b. Goals and Responsibilities
2. Data and Transparency Panel Activities to Date a. CDO Forum, Presentations from other states and signing of the Public Safety Interagency Data Sharing Agreement b. Gap Analysis c. State Data Catalog d. Feasibility and Cost Study on the Development of a Statewide Data Warehouse Program (DRAFT) e. Proof of Concepts
3. End Goal a. Centralized Data Hub for effective, governed data sharing, data-driven decision making, and coordination of services b. Sustainable funding model for centralized data sharing and data-driven decision support services
4. Current Barriers a. Data Sharing b. Funding
Other Business
Adjournment
Documents
| Title | Type | Pages | Source |
|---|---|---|---|
| Agenda — DATA SHARING & DATA-DRIVEN DECISION-MAKING TASK FORCE, Jul 18, 2019 | Agenda | 2 | Official source ↗ |
| Data Sharing Approved Minutes | Exhibit | 2 | Official source ↗ |
| Data Sharing Draft Minutes 7 18 19 | Exhibit | 2 | Official source ↗ |
| Handout 1 | Exhibit | 1 needs OCR | Official source ↗ |
| Handout 2 | Exhibit | 11 needs OCR | Official source ↗ |
Speakers
Speaker 1
Senator Jane English
Unverified
Jessica Jones
Unverified
Speaker 15
Speaker 16
Speaker 18
Adita Kerkera
Unverified
Speaker 26
Speaker 42
Representative Austin McCollum
Unverified
Representative Fredrick J. Love
Unverified
Senator Kim Hammer
Unverified
Senator Will Bond
Unverified
Speaker 80
Speaker 82
Senator James Sturch
Unverified
Speaker 21
Speaker 97
Senator Breanne Davis
Unverified
Representative Andy Davis
Unverified
Speaker 111
Speaker 117
Speaker 79
Speaker 121
Speaker 131
Speaker 135
Speaker 143
Speaker 91
Speaker 95