Inspiration

The different Swedish governments agencies have a variety of data open for the public. They are however not aggregated in the same place, but rather the different institutions have their own data access points. This makes it difficult to gather and compare data from several actors, which have the possibility to give new insight into the state of Sweden and its population. Our mission was to combine different data to see wether any correlation could be found between health data and other factors in order to indentify and visualize health beneficial factors.

What it does

The application "Korrelationskollen" gathers data from Socialstyrelsen and Statistiska Centralbyrån where the former provides causes of death and the latter factors like unemployment and economic fortune. By choosing two factors, the data is averaged for each region (län) in Sweden and presented in a graph for easy visualization. The correlation between the two dataseries is calculated and presented as well.

How we build it

With a strong backend knowledge base, the focus was put on building an API that could be used with different user interfaces, or as input to other applications. We used a python server coupled with a simple javascript app.

Challenges we run into

The main challenge was to find relevant data, that was both accessible in a managable format and had the right geographical resolution (Sweden's 21 regions in our case).

Accomplishments

The application is functioning prototype and can be used as intended with more than 20 parameters.

Features

• Choosing two parameters from a drop-down menu
• Instantly plots the two categories in a graph
• When a changing a parameter, the view is asynchronically updated
• Correlation value between the parameters is clearly displayed
• Each datapoint is for a certain region and which region is shown when hovering over the points in the graph
• SCB data for other parameters can easily be added to the project
• All available data categories for cause of death from Socialstyrelsen are automatically detected and added to the drop-down and will thus dynamically update if changed at Socialstyrelsen. (We used aggregated categories but up to 1900 parameters at the moment can be found there.)

What we learned

We learned the potential in looking at data from different sources, since even though the governmental agencies have their separate subject areas, there are most likely correlations to be found and evaluated!

Whats next

• Include more agencies and more data factors.
• Use more data for the correlation calculations (match time periods)
• Extend the functionality to handle data for each municipality in Sweden for example.
• Map visualizations.

How to Run

• Go to the root directory and run: $ python3 __main__.py
• This starts the webserver and now you can run the web application locally by browsing to localhost:8080.
• The api is available at localhost:8080/scb<Unemployment|Fortune> and /soc
• DiagnoseID's can be found at: http://sdb.socialstyrelsen.se/api/v1/sv/dodsorsaker/diagnos

TEAM MUMEÅ

Nimer

Backend
CSS
Architecture

Louise

Backend
Frontend
Statistics

Miriam

Frontend
API

Fredrik

Backend
API

Emilia

Backend
API