Visit: Shelter Associates Office
Date of visit: 17th January 2018
Location: Pune, Maharashtra, India
Purpose of Visit:
To understand the data collection process and understand the issues faced during use of software tools.
About Shelter Associates:
Shelter Associates (SA) is helping to improve sanitation in urban poor areas/slums. Data collection tools are a major driving factor for such projects. They have been recipients of awards such as “Google Earth Hero” in the social sector for mapping poverty, NASSCOM Social Innovation Forum Award, HUDCO Award for best practices in Sanitation being a few reputed ones.
The organization uses KoboToolbox. Kobo is a data collection suite which is part of the OpenRosa ecosystem that ODK belongs to and so there is similar functionality. For example, Kobo Collect is a fork of ODK Collect.
SA has written a great blog post ([Making the complex simple])(http://shelter-associates.org/blog/making-the-complex-simple-kobo-toolbox-improves-efficiency-in-data-collection/) on their use of Kobo that you should read! The tools they use are:
During my visit, I interacted with the following persons:
Smitha Kale - Role M&E Manager
Anova S Deve - Role Operations Research Analyst
Abhijit Ghate - Role Software engineer
Amar Kamthe - Role Consultant Software engineer
I demonstrated ODK Collect, ODK Build and ODK Aggregate and some of the features and we had a discussion about their use for data collection.
As the organization uses Kobo for data collection, it was a barrier to understand the issues from technical aspects as many of them were specifically related to Kobo. But, on the other side, many issues were also relatable to ODK tools.
During the discussion, there were some issues faced by data managers. Here is a list of some of the important points that I noted down that might be helpful for the ODK community to consider as important features.
Data visualization tool
The organization has built a custom data visualization tool to plot the collected data
Question Library makes form designer easier
While building a form, form designer can add the question from a library which helps to reduce the task of redesigning the complex questions again.
What is the use of geo trace for Shelter Associates?
Surveys of interior spaces in densely built environments such as slums using a mobile have not proved to be optimal for the application.
How data collectors currently map area of land?
Kobo collect is currently missing the geotrace widget for mapping. (Correct me if I am wrong)
How can ODK collect help? (Features of ODK geo trace)
ODK collect can be used to trace a polygon on a map. There are widgets related to mapping like:
- Geo point
- Geo trace
- Geo shape
Issues in bulk edit of data on server side
Currently, the data managers face the issue of editing data in bulk. They expect a tool to edit multiple entries of data a click of a button. For example, if a data collector entered a field incorrectly for all the forms, it should be editable easily.
The questions created in the questionnaire does not have a serial number in the interface.
Filter and download data
Data managers requested for a feature on the server side to selectively filter data and download it. Currently, there is no option to apply filters. Complete dataset has to be downloaded.
Renaming/naming of photos and downloading photos with filters.
The pictures collected during survey play an important role in the organization. Currently, there is no option in the data collection app to name the picture taken with the camera. It is important as it helps to easily find the required image manually.
Feature request: Name widget
The organization uses a form where data collector needs to fill an entry with a name (e.g., Mr. Xyz Abc). According to data managers, it is expected that data has to be entered in some specific format. Like CamelCasing, the language of filling and adding certain abbreviations like Mr.
But, data collectors fail to follow this process. So, it was requested that a custom widget for filling name would be helpful and save time for data cleaning and correction.
Should be able to choose abbreviation from a dropdown list which is specified during form designing.
Should be able to enter the name in First name, Middle name, Last name format if specified during form designing.
Follow CamelCasing if specified.
Feature request: Currency widget
Data collectors need to fill data regarding construction costs. Many data collectors enter the amount in the form Rs. 123, rs 123 ,etc. (Rs. stands for INR). Data collected in such form adds an overhead for data managers to clean it. It was requested that a custom screen for entering the currency would help to make happen data collection easier as the organization frequently requires currency related data during form filling.
A logo of currency should be displayed on the screen with a text field to enter the amount.
Difficulty in managing versions
According to data managers, when a new version of form was pushed to the Kobo server, the data collected from old form version was lost. I was unsure if this was a bug or a misconfiguration, but I would like to report it to the Kobo team.
Due to my limited knowledge about Aggregate, I couldn’t provide much information about how Aggregate manages versions. But, according to Hélène Martin (@ln), “Aggregate allows for some changes to the form without creating a new one. See http://docs.opendatakit.org/form-update/ for details. If the structure of the form changes, a new form needs to be created.”
Unfortunately, there were no data collectors available in the office. But, one of the major projects taken by the organization has been carried out in Sangli, Maharashtra, India. (Coincidently, it is my hometown.)
Overall, it was a great experience to interact people using data collection tools. It has truly inspired me to dive deeper into technical aspects and help to make data collection easier. Data collection tools are helping humanitarian projects to be carried smoothly. A lot of improvements can be made on the technical side. Data collected through tools play a vital role in decision making and for this to happen it is aimed to improve data collection process.
[From left: Me (Akshay Patel), Anova S Deve, Intern, Smitha Kale, person 5, Amar Kamthe, Abhijit Ghate]