Innovating at the Nexus of Big Data and Energy Access

Innovating at the Nexus of Big Data and Energy Access

Panelists: Kate Steel (Google),Lesley Marincola (Angaza Design), Michael Nique (GSMA)
Moderation: Peter Alstone (Energy and Resources Group, UC Berkeley)

Rapporteur: Hill and Katie McCloskey

Issues Presented

Michael Nique (GSMA)

• Mobile phones are currently an incredible resource—but do not confuse # of phones with # of sim cards (estimated avg. 2 sim cards per person).

1. Define big data in the energy space, and how does your organization relate?

• The plethora of phone data
• GGSMA can help identify means by which operators can collect end user data on fundamental calls, airtime, location that can be used for credit scoring in establishing the credibility of folks to attain loans.
• An example of the usefulness of big data. In the Ivory Coast, city planners were able to use phone data to identify new, useful bus lines as the popularity of various regions of their cities ebbed and flowed (they were noticing that folks were taking previously unidentified routes that required multiple bus transfers)

2. What are barriers to developing big data?

• Security and privacy
• Processes that allow dissection of copious amounts of various types of data

Kate Steel (Google)

1. Define big data in the energy space, and how does your organization relate?

• Google is a major investor (half a billion dollars) in clean energy around the globe. They have shifted to emerging markets—especially E. Africa and S. Asia. They also work with bottom-up energy access models.
• Google is interested in helping to parse through smart meter data.

2. Sources of big data?

• Example. If you could monitor Tweets to see where blackouts occur… geographically pin-point high-frequency black out regions.
• If you could look at a country, know where the population is distributed, what sort of credit do they have, how reliable are they for loans, etc., many of the world’s energy problems could be solved much more quickly.

3. What are barriers to developing big data?

• Processes: the types of data we have don’t exactly translate into useful information—e.g. knowing someone topped up their phone 50 shillings this month isn’t completely useful. But that information can be made useful

Lesley Marincola (Angaza Design)

1. Define big data in the energy space, and how does your organization relate?

• Work in the pay-as-you-go solar space, and are very interested in two-way data transfer—install it w/every unit they sell. They also provide user interfaces that help visualize clean energy usage, help w/after sales support to ensure highly efficient operation by the end-user.

2. What challenges can we overcome with big datasets that we couldn’t overcome with the small data we had before?

• Airtime data

3. What are barriers to developing big data?

• Cost of hardware/two-way-data collection. We can’t afford to collect data on cheaper appliances yet (e.g. 30 watt solar lanterns), because it will substaintially increase cost and decrease access.
• Are the data valuable enough to merit this cost?

Q & A

1. How do we use these data smartly? We have much knowledge, but not enough wisdom.

2. Who is currently using the data the most?

• Microfinance folks.
• Michael: phone providers should partner with orgs that have incentive to leverage this data (e.g. microfinancers, clean water orgs, etc.)

3. Do you work with the Global South so that their nations, businesses, and end users can have access to their own data for help with expansion of the grid?

• Lesley: Angaza does not work with governments. They do not feel that expanding the grid will solve energy access problems by itself.
• Michael: GSMA does tend to work with utility organizations/roviders

4. Will cost of getting data ever exceed cost of the hardware?

• Lesley: yes—sometimes it already does (e.g. putting a GSM chip on a cheap solar lantern). Keep in mind that there are two types of costs: transaction costs and hardware costs.

5. What is the demand of data at a local base?

• Kate: Data is only useful if you can monetize or provide value to the user… and it still remains to be seen what is the financial value.

6. Which efforts currently exist to collect and streamline these big datasets?

• Peter Alstone: Very hard to think about how to link datasets before you have a question in mind… it’s hard to know what to do first if you’re just waiting around for something to materialize.