We're publishing the article that our CoopDevs colleagues shared on their website, as they are responsible for creating the collective reporting channel forms, as well as the database that allows for the analysis, processing, and visualization of all received cases.
The original article can be read here: https://coopdevs.coop/es/blog/el-blog-de-coopdevs-1/business-intelligence-un-cas-practic-ovim-47
Context
Observatori de Violències Institucionals Masclistes (OVIM) (https://ovim.org/ca/) is a platform of civil society, of political and social influence, that sees why institutions act with careful diligence and guarantee the right to a life free of domestic violence and discrimination against children, children and adolescents.
The entities that have promoted this platform are: Almena feminist cooperative (https://almenafeminista.org/) i Hèlia (https://heliadones.org/)
On the platform, those who have participated in masclist institutional violence, their environment, or professional or activist organizations, can register their case, anonymously.
A cop registers the case, the team of Almena and Hèlia is responsible for compiling another, more extensive, formulary, classifying each case into different indicators, according to the information they have introduced in the report.
A cop classificats els relats es volen poder extreure statistics sobre la typologia de les violències institutionales masclistes recorded.
Execució del projecte
In this project, in addition to the data analysis part, they have also been able to generate two forms, the introduction of data by the people and the analysis. It was very important that these forms, especially the introduction of information by people, had the main aspect that remains of the website.
These formulas are based on django and are deployed to the main server in the analytical data base. The deployment has been designed to automate the analytical server functions through the ansible provisioning system that creates the entire business intelligence infrastructure (database, ETLs with DBT, Airflow, Superset, etc.).
A copy of the data to the database of the django application, the data have been mapped with the Foreign Data Wrapper to the datawharehouse and the data have been persisted with DBT.
A special feature of the OVIM database is that it contains only a few files (some hundreds of cases), but it also has several different columns to analyze the data. There are also many many-to-many relations, which by denormalizing them based on facts, cause the file names to increase. The fact of having several different columns is of particular importance in the definition of the Superset datasets that are only present in each dimension, so that they remain easily locatable when displaying the data.
A definition of the datasets in the Superset is that it will be possible to create different dashboards to analyze the data, which have served, among other things, to generate the annual report on major institutional violence for 2024 (https://ovim.org/wp-content/uploads/OVIM_Informe_CAT_v3_web.pdf ).
For the work of defining the data base and the Superset datasets, a very constant work has been necessary for the German and Helian team, since there are no studies on VIM, and it is necessary to define the vigilant process of analyzing the maximum possible variables, simplifying to the maximum the process for the people who have to enter them. facts. It is also possible to annotate the fields of the dataset in the most comprehensive way possible, so that Almena and Hèlia can create their own analyses.
Següents stepsos
The data that is required to analyze major institutional violations, especially in the first months of the project's visa, go to the extent that it is fan of our analysis, in addition to modifying the analytical form and the Superset datasets, always maintaining also the original structure to avoid the loss of data. How to maintain the project so that it remains easy to add this new information, both in the data entry and in the visualization.
In the last part of the project they have defined some graphics to the public website (https://ovim.org/ca/datos/), but it faces only a few. The future objective is to be able to create a dynamic visualization, via the web, that shows and explains the facts of major institutional violence.
As always, you can check all the code used in our Gitlab: https://git.coopdevs.org/coopdevs/bi/ovim
