 Okay, we're gonna get started with the next session. Welcome to all the new people in the room. I'm Danielle, and I'm introducing the sessions in this awesome session about cities and datas. And next, we have Gabba Rodriguez, and she's an activist and a hacker who works at the Coral Project and does consulting around data. She loves the intersection between media and technology. She co-founded a nonprofit called DATA, which works with open data and transparency in South America. She grew up in Uruguay, and now she hangs out with me in Portland. So, Gabba. Thank you. Hello. Hi. Hi, everybody. See me? Okay. So, today I want to talk a little about a project that I was involved until last year. I'm a software engineer, but I also work with data with different nonprofits on different governments. So, I help Mexico City to implement a standard on open data. So, I want to talk a little about my experience there and also about what is the standard. Let me move it because I really need to... It's better this way. Yeah, it's still there. Okay. So, first, what are public contracts? So, something else, two disclaimers. One, English is my second language, so I'm sorry if there's any mistakes. And second, I'm sorry there are no cats. So, yeah. What are public contracts? So, an agreement to perform a particular task to benefit the community at large that is financed by government funds. Those could be from like building of bridges or airports like in Mexico City to contracts of cleaning services in the different organism. So, those are public contracts. Why is it important to open them? Why we need... This is very... It's quite related with the talk before me, seems to me because it's one of the reasons why we need to open data, no? In the case of public contracts, it's most of the corruption that we have in our governments is through public contracts. So, having open data makes them transparent for everybody to use them, to cross them with other data, to see where there could be problems, allow to track the tenders and awards, allow the government to monitor implementation on how the public contracts deliver services, deliver a better value for money for the government. And that is a very important reason on why to open data around public contracts. It's not just for the companies that are applying to these contracts to benefits from where the contracts are, but also for the government to be able to know what they are spending money on, no? That usually, as she was saying before, is usually different dependencies do not know what the other data and other dependencies are about, and so there's not much information inside the government. So, opening the data makes the civil society to work on that and help the government to deliver better value. And of course, prevent fraud and corruption as the first one to reason to open public contracts. So, in around 2013 to 2014, there was a lot of interest in pushing different governments to open in contracts. So there was a partnership that was born, that's called the Open Contracting Partnership. There were like policy expert, programmers, leaders, campaigners that tried to see how that could happen, how they can build a standard on open data. That is not only about the format and meaning of the data, but also about how to publish data, how a guide on from the beginning, because it's not as simple as saying, okay, open the data. Sometimes the government don't have data on the contracts or everything's in paper like in the case of Mexico. So these partnerships try to advocate around opening public contracts. They open public contract through disclosure, data and engagement, so that a huge sums of money involved are spent honestly, fairly and effectively. So the contracting process. So this is in general, around the world for all the governments. We have different phases when contracting. We begin in the planning, so when the different organizations try to think, okay, this is what we need, this is the budget we have, these are the project plans, this is a market to study, so there's a lot of documents and information related where the government wants to spend the money. After they have a specific plan, they go and post a tender, so they have notices to call the companies to say, okay, these are the tenders that you can apply to. So in that case, there's documents around tender notices, specifications, inquiries, and then there's different processes, depending mostly on the governments, on how they award that contract, no? In that case, include the tales of award, the Peter information, Peter evaluation, and then when there's a decision made about who gets the contract, then you have to sign the contract. There's the sign contracts, amendments, if there's some changes, finally tales, and also sometimes information about contracts that do not expire, that renew automatically, there's many ways that the contract could continue. And then the implementation, that is the part that mostly the governments don't have so much, but it's quite important that it's about monitoring how that contract went through, all the milestones, the payments, and the termination info that's appropriate. So from there, from that process, we build what we call open contracting data standard, that data standards, as you may or may not know, are the rules by which data is described and recorded. It could be standardized the format, as well as the meaning. You're saying, okay, you have to publish, see the Shazons or an API or something like that, and this is the information you need for each phase. And the specific open contracted data standard is a global non-properitary standard structured to reflect the complete contracting cycle. So it tells you how to publish data in each phase. How periodically to publish the data. The approach that we had around the standard is to publish early and iterate. So for example, most of the government can publish documents. That would be the first part, no? Not published structure data, but just the documents and the contracts, that would be something, the first step to start publishing data around open contracting. Then the other approach is try to have them to publish in a Shazons structure and publish data for each step of the contracting process, create a summary record for an overall contracting process, re-use objects, so for Mexico, we try to get the list of organizations, standard information, line items, milestones, different kind of information around contractings in general, no? Then recommend data and document at basic, intermediate and advanced level. So when we try to approach about opening contracts, we say, okay, you don't need to just do the five stars of open data. I can talk later about that, but you just publish the basic information, whatever you have, and then from there, well, now try to add this, now until we have all the levels of what people may use or may need. And the other approach is guidance on improving data collection and data quality, no? And a growing community of users on range of open source tools. Something very interesting about the open contracted data standard, for me at least, it's quite interesting, is that it's a community effort, so it's working the open, so the standard is being developed in GitHub, and you, as whoever can go and see the different versions and see what are the possible extensions and possible features for the next version, and this is all filled from the specific implementations, implementations in Mexico, in Ukraine, in Paraguay, in different places, are feeding back to the standard on which things they need or don't need, or what things could be good or bad about how it's been implemented, no? So, first, I'm going to go through different phases of how we go about implementing in Mexico, how it's been implemented in general around the world, no? First, we do an assessment, and we ask these questions, no? What data systems do we have that hold contracting data? That is the harder part about implementing this, because sometimes the information for different phases are like all over the place in different systems, so we need to find a way to extract all the data from different places and publish it. For the case of Mexico, the Mexico City, no? Because the federal level is more complicated, and there's also an effort about implementing it at the federal level. For Mexico City, we found a specific organism, the Secretaria de Finanza, that holds most of the information about contracts, so they have the technical capacity to implement the standard, so we work with them. And from there, we change their system to be able to publish data the way we were recommended, no? The second question would be, which data system hold data on each stage of a contracting process, no? That's, yeah. Like, for example, for the case of implementation, it could be harder, it could be in some other systems. Do you hold, something else, the other question is, do you hold consistent identifiers for each contracting process and for the parties to tender on contracts? So this is an effort that tries to bring public contracts around the world. So what we try to do is have a specific, unique identifier per contract that is global. So for Mexico, they have the identifiers, we give like a code, and with that code, they made the identifier, and that says all the contracts that start with that code are from Mexico. So people could, for example, like journalists or civil society could interact with contracts from different places, and they could understand also how that works, no? And then the last question is, what technical resources do you have available to help in implementation of open contracting that is quite important to, and in the case of Mexico City, yeah, we had a team that was quite willing to help with implementation. So here is like, continue with assessments. We write which systems they have for each of the stages, planning to end their work on their implementation. We ask, is it data as a structure? For the case of Mexico City, they had a law, a transparency law, that says, for whole Mexico, that says every three months, they need to publish a contract, but what they were doing is that somebody was going through all the papers and adding to a spreadsheet manually, every three months, all the contracts, and they were published that. So there was some information, but it was not coming automatically from their systems. So yeah, then we ask if the documents are published, that would be the first recommendation when it started implementing, because that's the validation of what we are seeing in the data, no? So that verifies that the data that they are publishing is correct. Then the revision history, we try to publish like photos of each part. You have the whole process, so we say, okay, first publish a photo of the planning, then when it goes to tender, a photo, so you have control of different history revisions of the process, and with that you can have a better idea how the contract move on. And then access restriction, because many governments, when they publish data, they require a login, for example, or an account. So one recommendation is that when they publish data, it should have no restrictions on access, no? So the Open Contracted Data Standard provides a structure data model for capturing information about each of the stages of the contracting process, and we have a JSON schema which describes all the field names, field definitions, and structures for the data, no? And then when we talk about compliance, we talk about assessing against that schema, no? So we talk about the JSON schema and also the guide on how to implement and publish the standard. So this would be, okay, so we had an assessment, now the next step will be a two map data. So for Mexico, we did that, we see, okay, this is the recommendation of the standard, no, we have this JSON file with all the different stages. So the first one, we fill planning and we send that with public information. The second release, we fill tender and we publish that information, and that's how you start filling it up, no? That's JSON file. And what we do to be able to implement this or to be changed systems or other new system that they have about publishing data is that we map, we say, okay, the planning is in this system, is this data, that's where it's coming from, that's how you call it in Mexico, no? And try to see rules about what's happening around planning or so there's a lot of things around mapping that that is the most hard part of this process. Then we publish data in machine readable releases and records, that would be the JSON files with a unique identifier. And then we publish using publication patterns. This is what I was talking about the five stars. So there's a kind of quality score, no, about how governments publish data. So first would be the first star, we give the first star, one star, I'm sorry, if they upload basic contracting data and documents to the web, just do that. So Mexico City already had one star because they were doing that. They were publishing the documents and they were publishing manually. Two stars, if they provide machine readable data, three stars if they use the standard, four stars if they provide an API access to the data so people can use that API to get data, and the five stars if they provide other kind of data sets that we can cross open the contracts with, no? Then the next step after the publishing and define all these stars is the check the validation of the data. We have, if you go to GitHub to the open contracting data standard organizations, we have a validator that does a technical check on the way the data is structured and format. Then we look at the conformance, so checking if what they are publishing and the concept of each terms is used correctly as the standard says, no? Then how complete is the information they are sending and then if the data fit for purpose or not, no? This is something that seems to me very interesting about this specific standard on the data that is like community built. This is, yes, is supported by the open contracting partnership but it's also built by all the different civil society organizations and government around the world that are trying to implement open contracts. So you go there and you can like in the standard repo look at all the issues and all the questions that comes from different implementations on which things we need to change and how we can extend this standard to make sense in the different places, no? And this is the extension. So the extension is, for example, I have this standard in the version one, no? Unlike Mexico says, okay, we need this, for example, this thing about contract. There are some contracts that never expire and they renew automatically. Okay, that was not in the standard in the version one, point zero. Okay, we need an extension. So they publish it the way they want, they propose the extension and then in the next release of the standard we see if that extension goes into the core or not. So that's how it works, the building. History of releasing data in Mexico City. So first, yeah, this would be the standard, no? Now we go back to Mexico. So they validated the standard in 2014 and was the first city to release is full contracting data from planning to implementation. So now they are releasing data already in the standard. There was a big thing about this is politics that was very hard. So there's a lot of politics about pushing it forward to implement the standard. Yeah, that happens in all the forms, I guess. Starting a small department with technical capacity and willing to change processes, that's something I already talked about. They have the technical capacity to implement the standard. Also they have the willing to change because it's not just to add a system and say, okay, publish data, but also the cultural process, on how they capture the data, which new things they need to come to. So there was a lot of internal training about how things need to change for the people that are capturing data that are sending it from paper to the system, no? And we try to involve and we recommend to involve the civil society from the beginning, so we started doing, because one important thing about open data is we need to see why we need that data. What are the cases? In the case of Mexico, people were really concerned about the metro change, concerned about airport, there were many things that people in Mexico were concerned about, so it was important to talk with them to see which kind of data they needed, no? And then the modified systems to capture needed data. This is the website that some people build from the data that the city publish, and this has an API and also has visualizations of how the different contracts happen, and different transactions. So up there in the URL, you can see the identifier, the unique global identifier from Oathman Contracting for that specific contract, no? Challenge, contracting process identifiers is sometimes, this is something that I already, I think I talk a little about already. The identifiers sometimes are not so easy because sometimes the systems have information here and there, so it's hard to match and create a unique identifier. Useful data, you start with this organization that has the technical capacity and willing to change, but sometimes it's data only about not very useful. So that was one challenge. When it's starting involving the civil society all the way, it's hard because mostly in Mexico and I guess in other places, it's hard for the civil society to talk with the government. Pushing to open critical organization process and culture and capturing data, something I already talk about. Other places implementing OCDS, very interesting if you really want to read about open contracting, is the case of Ukraine. There was a civil society group that started pushing to open contracts and they build this system called Prosoro that is free software and the government implement Prosoro in their own organism. So now they have this website that publish contracts. And other interesting places, Paraguay too, that this is a visualization that they did. For example, there, they see what are the providers? Yes, they are the providers that get more contracts from the government. So it's interesting by time. So if you go to their website, you can see all the interesting visualizations. That's it. I will start the links if you're interested about this. Thank you.