Friday, June 24, 2016

Mapping Open Data in Armenia

The accumulated and released open data of many state institutions make up quite large volumes in Armenia today. There are already a few vivid examples when the use of open data has had a positive effect.
A number of databases have large potential for anti-corruption and civic journalistic investigations; for example, "Purchases made by one person," "SNCO Financing," "High-ranking officials' declarations." These are examples when databases allow a researcher to carry out independent research work.
But it's also important for the data to be subjected to automated processing as much as possible, as well as presented visually. Thus, for example, the Swedish government provides a special API to receive and process statistical data in an automated fashion, while the US government provides a separate large tool through which visualization of open data is possible.
In Armenia's case, there is one publicly known data visualization tool: the interactive budget, which doesn't provide the user with many opportunities for modifications.
As for the opportunities for processing data, Armenia has not yet obtained great victories in this matter. The National Statistical Service provides databases with very minimalistic tools. Apart from this, there are an additional two platforms for processing statistical data: Armstatbank and Armdevinfo. Both are quite difficult and inconvenient to work with, and are not modern.
On the other hand, there are specific cases when after their automated processing, state data are provided in a much more accessible appearance. The official databases on taxes and the Harkatu ["taxpayer"] platform created by a private company to develop them can be compared.
There is, however, a case having a far greater influence, one connected to currency exchange, gathering in one place the currency exchange rates declared by all banks in Armenia, had an effect on the entire market, bringing currency exchange rates around the country to approximately the same level.
In open data, several components are immediately important. Data must be
  • Accessible
  • Complete
  • Available through archives
  • In the same format
  • In a format ready to be developed
But in Armenia's case, problems arise connected to access to the archives. After websites are updated and modernized, there are instances when the archives disappear. Thus, for example, it's no longer possible to find information connected to previous administrations on the Armenian president's website.
On the uniformity of data provision, there is the government decision "On approval of the minimum requirements for the official websites of state bodies online" [AM]. According to this decision, the website must adhere to general norms. But the reality is the state part of the internet is too diverse, confusing, and in some sections has a structure that is illogical and difficult to understand.
There are also different approaches on providing the data. According to paragraph 5 of the aforementioned decision, information posted on an official website must be complete and free of charge. The matter of completeness can always be debatable.
The Electronic Register site, for instance, requires a payment of 3,000 dram [about $6 USD] for complete information on LLCs. For each inquiry. That is to say, in this case, the data is viewed not from the perspective of freedom of information, but as a provided service.
The strong diversity of the ways of providing open databases and their being decentralized today is an obstacle for analyzing and processing data, which complicates applied approaches.

Current open data databases in Armenia, on one hand, provide quite a lot of information, but, on the other hand, the information is provided without a specific format, a clear form. The way information is provided changes from site to site. On one site, it's simply in html format; in another, in .pdf, .doc, or .xls format.
There are also databases that provide data only through search. There isn't a common approach on how and in what format the information may be published. The legislative requirement, in fact, is being met, the data is being published; however, working with the data becomes difficult.
One can come across data that is already outdated or inaccurate or is technically not available on the site.
The work becomes difficult not only for a specific exploration — in many cases, automated work with the data becomes almost impossible.
The information provided by a state agency, when adapted by a third party, can become grounds for numerous programs to improve public life: for example, systematic and automatically processed information of the education system, transportation, environment, agriculture, and other sectors. But to find such solutions, the information must
  • always be recent, reviewed, precise;
  • be in a machine-readable format; for example, XML, CSV, and so on; and
  • be provided in a few formats.
Special API, which is still neglected in Armenia today, becomes important in working with information today.
Numerous other questions arise regarding the presented data. For example, the approach to ownership is quite inaccessible. The copyright symbol and the corresponding text "© All rights reserved" can be found on all the government websites (only on the president's official website is the Creative Commons 3.0 license also posted, which implies that the content is more free to use. Though, in this case, it's unclear why the 3.0 version was selected).
This approach means restricting data processing and republishing. On the other hand, the vast majority of the information posted on government websites — the news, databases, documents — are generally not subject to copyright. And those few pieces that might be subject to copyright (for example, photos and videos) were created with funds from the state budget, and copyright restrictions on these run contrary to the public interest.

It's understandable that the copyright symbol is automatically placed on government websites, imitating the overall approach online. But the uncoordinated approach to such issues provides unscrupulous officials with the opportunity for abuse in working with the information.

The list of databases with types of data and exploring possibilities you can find here

There are two big clusters of sites: Regional governments and Diplomatic Missions of Armenia
(you can find clusters here at the bottom of doc)

There are several clusters of databases: