Online collaboration by Nepal's data community yields a portal for opening local election data

We recently became aware of an exciting new project being conducted by Open Knowledge Nepal in collaboration with Code for Nepal in relation to the upcoming local elections. We spoke to Open Knowledge Nepal’s Nikesh Balami to find out more.

"I was pretty excited when the local election was announced - partly because it is an important milestone in Nepal's local governance landscape, but mainly because of the opportunities for interesting projects around opening up election data” said Nikesh. Nikesh has been involved in various open data projects such as opening up national budget, opening up local government spending data, and assessing the openness of government data in Nepal. Having worked with election data during the Constituent Assembly Election three years ago Nikesh and his team were enthused when the local election was announced recently. They decided that they wanted to do something related to Nepal’s local elections in May. He reached out to the online open data community via twitter to seek and share ideas for what could be done. 

He received a very positive response - "within three days 15 volunteers from outside and inside Nepal showed interest and were already exchanging ideas".  Within two weeks the online collaboration came to fruition and the sharing of data, coding and content led to the development of Election Nepal - a portal that shares all kinds of data on the local elections. Nikesh shared that "it was awesome to experience the process of building a product from scratch solely through online collaborative work-without having to meet-up physically. This demonstrates the potential of the online data community and crowdsourcing".  

Nikesh explained that the aim of the portal is to "present local election data in an accessible way so that even people with low digital data literacy can use the information to inform their voting decisions". Nikesh shared that "election data and information published by government sources are mostly less categorised which makes understanding data difficult and time consuming. The data is scattered and published in formats such as PDF and JPEG which makes finding, accessing and using them difficult". To address these challenges the online data community worked to integrate the election information and data from various sources into one common platform. The portal (currently published in Beta) has three areas of focus - visualizations, datasets, and the twitter feed.

i) The visualization section has interactive visualizations that help viewers to find out about the division of federal state structures and the number of local bodies in each district. It also includes infographics on the Election code of Conduct and the process of voting. 

ii) The datasets section shares data such as number of political parties in each district, number of local bodies, election areas, results and candidates. This data has been obtained from various government sources and shared in CSV format to enable its reuse.

iii) The twitter feed is included to highlight the social media buzz around the local elections and encourage interaction with the data.

A shared enthusiasm for data has helped the online team work together to set up this live portal, but they have had their share of challenges too.

         i.            Because of the unavailability of geo-coded data for the new federal structure the team was not able to create interactive maps demarking the new levels of local administrative units. For example, there was no geo-coded data to map how much of a district is covered by a metropolitan city or an office of the rural municipal executive (levels of local administrative units under the new federal structure).

       ii.            The data and information in the Election Commission's online publications consists of scanned images (as opposed to the more commonly used PDF format) which made it impossible to scrape data. The data from the images had to be manually entered which depleted a lot of the team’s available time.

      iii.            The publications of the Election Commission use the Nepali font PREETI. Because of this the content got distorted during the process of converting the text into UNICODE. Cleaning the data took the team a lot of time as they had to refer back to the original publications to correct the distorted data.

     iv.            Before the Election Commission started publishing data on the number of local administrative units in each district they had to refer to various online portals. Different online portals published different numbers at different times which added extra hours to the team’s work as they had to validate the data.

The team is now working to find alternative ways to deal with these challenges, continuously working to improve the features of the portal and update it with the latest datasets. They are also working to create new visualizations which they plan to publish soon in local newspapers. If you are passionate about working with data and interested in joining the online team, sharing your ideas or contributing code to this project please get in touch!