When working with data, visualizations come in handy in answering questions as well as assisting in formulation of queries alike.

The process of creating data visualizations can be daunting but breaking it down into smaller processes makes it easier and manageable. Lets have a take.


Get to establish the kind of data you require and whether its from a file, a network or over the internet.

Data is collected a vast no of resources such as as books, pdfs, worksheets, log files for databases.

For the exercise I’ll use data from IEBC-Kenya. This is a case where thousands of persons of shared identification documents and a was a published a story in January. The data I got was published by IEBC on their website found here:


The data published is in a pdf and which is not machine readable. There are a variety of ways that the tables could be extracted and for this case I used Tabula which is an open-source tool and easy to use.

once you upload the pdf to Tabula, select the table you need information from and choose the format the data is to be extracted in. The data extracted from the pdf comes in forms of CSV, TSV, JSON & SCRIPT. (for now i opt for CSV and export it)

To give the the data some meaning, have it in a spreadsheet to allow ordering it into various categories which makes it easier while working on the data in our next step.

Clean it.

Not all the data you extract will be of interest to your story and as such remove all redundant point.

Filter out those that are not the point of focus after all the aim is to seek insight that data is speaking.

Map it !

There are a couple of ways that you could present your data ranging from charts, infographs, timelines. For this as it pertains distribution of individuals across the country , I Choose a visual map to show it. You can create a free account with cartoDB and upload your data(get the shapefiles here).

Now that the data is represented on a map, refine it with legends, hovering options highlight how the data could clearly bring out the stories lastly and onto the last step ,


With a good representation, the kind of story that the data is telling should be clear and visually engaging.

When this is the case the right queries can be made so that the data is relied on expounding on the issues highlighted.
Carefully scrutinize it as the possibilities are countless.