September 9, 2012

Thoughts from NDI Conference: Open Data & Elections

Letter dice with blue and white color and open text and open lock icon

This past week I was invited by the National Democratic Institute team in Kenya to speak at a conference they organised in Nairobo to discuss the elections. They asked me to discuss lessons from the Kenya Open Data Initiative that might be useful in thinking about elections management in Kenya and other countries. In the last few months, I have been grappling with this one subject and in recent days, I have only been thinking about one subject: relevance. And so that is the one thing that I found I have the capacity to discuss at the conference.

At the NDI Conference, I shared two essential ideas:

  • Relevance of Data is the Ultimate Driver of Usage

When I first got involved in the Open Data, the vision was that it would be available for every citizen, so that it would empower them to improve their lives. For me, that vision has not changed. However, I have refined in my mind what it would take to get that citizen to even have a desire for the data.

Data can only gain traction with the citizenry if it is useful from day to day. In the west, most successful data projects have been those that the citizens use for every day activities – transit, the health certification of restaurants, etc.

Data for development is a tougher proposition especially in Africa.transit or health restaurant data may not be as useful in Africa – especially when you consider some of our favorite places to eat and the informal (read: chaotic) nature of our public transport systems. It has dependencies that we must consider: literacy levels, exposure levels and interest. When thinking about education data we have to spend time considering what people want: often it has nothing really to do with teacher-student ratios. Rather it is likely to have more to do with space availability and past performance.

When thinking about taxes, citizens have more of an interest in how much they have to pay and what relief areas they have access to – including loopholes they can use to avoid paying taxes altogether. When they know what taxes they must pay, then they want to know what it does – “how much do I pay for the kenya police service? how much of my taxes went into planning the election?” Again, this is NOT the majority of the citizens. The majority are busy trying to move their lives forward.

  • Data relevance is increased if there is an interesting story around it.

Citizens are net consumers of knowledge and information – they have no major interest in analysing stuff for themselves. Take a look at how many people actually analyse their budgets beyond the higher level points. Check how many people actually have budgets for themselves period.

As net consumers of knowledge and information, citizens often only find relevance in data when that data is enveloped around interesting stories that relate to their lives. As a result, you may infer that data apps are not necessarily useful in themselves unless:

  1. they are helpful in the every day living – they help me work, move or live better/faster/ easier
  2. they are coupled with stories from analysts who tease the relevance of data to every day lives.

This isn’t to say that citizens are dumb. It is to say that they are too busy trying to make their lives easier and better to spend any time or brain power trying to understand data that does not necessarily change their lives today. Open Data creates security and secureness by it very existence and it helps citizens to feel comfortable even when they may not interact with it often directly. They at least use it to verify what the infomediaries (leaders, CSOs, media) tell them.

So therefore on elections?

The foregoing means that as we examine the role of Open Data for elections, we must think about the following:

  1. What exactly is the citizen interested in? (and this will vary from country to country)
    • Could she really be interested in voting patterns by location? To what extent is she interested in how much money was spent? (note that a majority of the citizens struggle with visualising what Millions and Billions are). To what extent is she interested in
  2. Who is interested in analysing and why?
    • if you are thinking about the citizen as opposed to the policy makers, then the analysis must be useful in making life easier such as engineering the election basis so that citizens go through it faster, easier, more comfortably. Te data that could be useful for this could be:
      • how much time have people needed to vote in the past? – by time of day, by demographics, by location?
      • How many average people need assistance, where, in how much time, at what time of day?
      • What stories can be shown?
      • How results are shared up the chain from the polling station level to the election management level to the public and how the are announced.

Just how Open must electoral institutions be with their data?

This came up at the conference. Someone asked me, what arguments should be made for electoral management agencies like Kenya’s IIEC to be open. In my view, the electoral commission should be cagey only on two aspect data – citizens private data and data that is crucial for the security of the elections. For example, they may not release the names of the people who will transport ballot papers or boxes before the elections. But they should do so after the elections. Everything relating to the process, how many people are available to count votes, where when, how many people turned up (by the minute updates) and voted; how many times they met with different candidates where and the discussions, etc. Every aspect of the elections that does not impact on the integrity of the process should be public.

At the end of my presentation, I launched into a “why not?” conversation: Blog on that coming up.

Share this post:
Facebook
Twitter
LinkedIn
WhatsApp

Discover more articles