Unlocking the potential of big data analytics: Mobile phone data workshop in Rwanda
Mobile phone data offers unique opportunities for national statistical offices in Africa. Without the necessary knowledge or skills, however, it is difficult for them to use this innovative data source for the benefit of their countries.
To help the national
statistical offices in Africa, the United Nations Statistics Division hosted a Workshop
on the Use of Mobile Phone Data alongside the 5th International Conference on
Big Data for Official Statistics. We spoke to Karoly Kovacs from the UN
Statistics Division about what knowledge was unlocked at the workshops and what
the future holds for African statistics in regard to the use of new data
We just had the workshop on mobile phone data. Can you tell us a bit about this workshop?
Karoly: We were here to organise two workshops, namely a workshop on scanner data for price statistics and a workshop on the use of mobile phone data – the latter was a little bit more widespread because the focus was not only on tourism statistics but also on human mobility. We wanted to convey the message that these new data sources could be used in the African countries, but they need preparation before starting any projects on the subject.
What made you think of holding this workshop, from the UN perspective?
Karoly: Actually, it was organised along with the 5th International Conference on Big Data for Official Statistics. The initiative came from Rwanda – the Rwandan government wanted to host this conference, and we thought that for Africa big data could be an alternative data source for making their statistics production faster and hopefully, with a greater breakdown. Later, the workshops could facilitate either starting new projects on the use of big data or boosting the existing ones.
Who attended this workshop?Karoly: We had 12 African countries represented at the workshop, plus Mongolia, Indonesia and Korea were also present. Many local participants came too, so the workshop was not small. We wanted to prove that one can just use a smaller data set even without significant knowledge or skills in big data and could make a good analysis of it. They can explore the data and see how to build a project and how to prepare themselves to step a little bit outside of the box.
Can you describe a little bit how this workshop ran?
Karoly: First of all, we had a little bit more IT people than statisticians in the room, at least from the international side, and for the locals, it was quite a good mixture of different government institutions. The agenda of the workshop had multiple goals. There was an international perspective where we started pointing out why this workshop would fit to bigger projects. We also introduced all possible data sources that participants could use, but we placed emphasis on the quality element, which is very important no matter whether new or traditional data sources or methods are used. We answered important questions like:
- Is this data useful for official statistics?
- How to integrate what they learn into their statistical system ?
There are many standards used in the area of statistics. They’re designed not for big data; however, they are guiding statisticians all over the world in their statistical work in different statistical areas. Along with these standards, whatever we do, whether we use traditional data or new data sources, the same quality aspects should be followed. I think that the message was conveyed very well.
During the workshop, we started reviewing mobile data sources and continued with access to mobile positioning data. We also ran a few data exercises in Excel, helping the participants to get some new skills and experience. Ms Tracey Li from Flowminder introduced several important aspects, like privacy and regulatory concerns, data access models, and partnership. Indonesia made a very interesting country case study showing how they used mobile positioning data for internal tourism purposes, underlining the benefits and drawbacks. Mr Siim Esko from Positium presented next on how to build the project up, which could be very crucial if a country is designing a new project based on MPD. Finally, we also had some discussion about the current research and the designed path for future projects on the last day.
Yes, it was to show that things have been done using mobile phone data, there is adequate support for these methods, including comparison with reference data, and there's also a lot of research out there.Karoly: I can support your statement. We also showcased some good existing, functioning country examples. The UN Global Working Group has an inventory as well as the UN Global Platform on their website. There are many different research papers or project papers available, which we shared with the participants.
How do you think the Global Platform and UN work could help African countries initiate mobile phone data projects?
Karoly: The examples I just talked about could help or whet their appetite but again, the countries, the national statistical offices need to figure out what they want. If it's clear mobile location data could be or would be available to use and would be beneficial in providing statistics quicker or making it possible to get greater breakdowns or other benefits, they need to make a project plan, focusing among other things on how to gain the necessary knowledge and/or necessary IT environment to execute the production of MPD-based statistics. And that's exactly what we wanted to show the countries present at the workshop, that it requires tedious preparation and planning first and data work comes later. I think it was useful to show that mobile phone data is not another statistical area where they just start a regular project, as it requires making a lot of preparations and acquiring skills before they can start such a project.
What do you think the African statistical offices and other statistical offices that participated in this workshop got out of it? What was the feedback?
Karoly: Let me start with the fact that there is an African perspective, which was quite well conveyed at the five different meetings during the week in Kigali, including the two workshops, a seminar, the meeting of the UN Global Working Group and the Conference. The UN Economic Commission for Africa (UN ECA) and the African Development Bank expressed that they are going to join forces and work together on developing projects on the use of big data, reflecting exactly what Africa wants. Because everybody is doing bits and pieces on the use of new data sources and they just realised how important it is to have coordination. I really think that this is already very good.
Getting back to your question, what is important is that the workshop gave the participants some help on how to set up a project. That includes a very strong recommendation to pick out what their starting reference points would be, so that they could decide for themselves which areas they would like to apply mobile data to. During the workshop countries talked about their present interest; however, it did not represent the official preference. Number two, it doesn't represent Africa either, given that only a few of all African countries were present at the workshop. That's why I think there should be some regional preference established but there should surely also be some serious support given to the countries training-wise as well as money-wise .
If they need additional training, what kind of training do you think would be the next step?
Karoly: I think before that they need to still make a kind of choice. Most probably not just for these 12 countries but they should make an assessment for Africa, and then we shall start preparing some more specific workshops. So, if the choice is truly tourism then we can focus on tourism work. But tourism is actually still not that important in Africa. I think transportation is more important and maybe migration because that's what we heard from quite a few countries. Population and migration are particularly important for some smaller countries. They would want to see how the population is changing. My gut feeling is that those topics would get a higher preference. But we'll have to see.
After the successful workshop in Kigali, the UN, BPS and Positium hosted another regional training for Asia in Jakarta. Between 11–14 June, ten national statistical offices came to the BPS head office for an intense four days of capacity building on mobile phone data.