Nowcasting for Green Urban Metabolism by Sustainably Innovate
Jun 19, 2014
Since you started your power company discussions in Berlin Germany - I say "Guten Morgen!" Your Big Data approach using data mining/metadata, SQL/syntax rules, etc. provides an excellent basis for business intelligence and analytics (BI/BA). Makes me think about my experiences here in the Washington DC area. My power provider is PEPCO here in Maryland. Yesterday we approached an all time high (record has been 97F) and I received a recorded message from PEPCO saying they are experiencing high power demands and that customers are encouraged to use power wisely. I am taking advantage of PEPCO's free service to track my power use by day, time and month, by providing a colored histogram chart. The PEPCO CIO team must be doing their due diligence in looking at the entire customer network and using regression/predictive modeling they can anticipate "expected" demands. So you are right about abnormalities affecting the prudent use of energy. As many data feeds as possible can help get wind of real changes coming. But there are some "locality norms" too. Here in DC for example, traffic is always lighter on Fridays since lots of people work flex hours (e.g., 10 hour days Mon-Thurs). Thanks for presenting these ideas. Good Luck on the project. Take a look at the gaming project here too. Very interesting approach and go to the Indian BEE website - really good stuff. Mark.
Jun 19, 2014
I was wondering what incentives can encourage power companies share data like that.
Jun 20, 2014
Greetings Pianpian! First time I've seen you post on the project's I'm looking at. Thank you. I see you have an environmental Master Degree from a NY university - great. Believe you'll agree power companies should embrace protecting the environment - basic good corporate citizenship and governance. Also, when commissions meet to approve or deny rate increases, the power companies should have empirically verified, validated energy distribution data for all to see. Thanks again! Mark
Jun 20, 2014
Dear S.I., Thanks for your proposal, very interesting. Tempting argument that it could work. Gaps challenges and next steps include the following: - how will the data be collected - privacy issues: opt-in? what are the incentives for opting in? - requires geo-locational devices and permissions? - who will develop the app? - how much does it cost? - who will fund it? - how will it make money? Serves as a sort of energy-efficiency technology for the utility? - proof that data regarding behavior in response to things like weather would allow utilities to save money/energy. - modeling to demonstrate that the savings provide cash flow adequate to cover development, financing and service costs. Good luck! Best, Jessica
Jun 20, 2014
Thank you for submitting this proposal. You identify a need for better forecasting abilities in order to provide cheaper/less carbon-intensive dispatch strategies and to better plan renewable energy technologies. You propose building better forecast models through mobile and social media sources. This is a very interesting idea. It would be helpful if you compared it with the current state of forecasting in a little more detail and explained how this approach would improve forecasting ability. Obviously at the proposal stage you don't have the full answers but I think some more work is required here. Other comments as follows: 1. Why Berlin? Why maybe the US? 2. Some discussion of current dispatch strategies in the location of interest, plus load profiles and other related data, would be helpful. It should be fairly easy to get at least some of this information for Berlin. 3. Again looking at Berlin (or German) data, you could estimate the annual carbon emissions from different sources of energy. This would provide an order-of-magnitude number for context. 4. In the summary section, I would re-work it so that the last paragraph is the first. Don't bury the lede! The example is helpful for readers to get a concrete example of what you are proposing, but I would see if it works better in a different section. 5. Who else is doing these kinds of analytics? Benchmarks or example cases would be interesting and might support your proposal. Best of luck, Mike
Jun 22, 2014
Hi S.I., Thank you for the proposal. Certainly worth exploring ways for utilities to dispatch their power plants more efficiently. I look forward to seeing some additional details added to this proposal and think that the comments already provided by the Catalysts provide useful direction in refining your proposal.
Aug 13, 2014
An interesting idea that is worth flushing out. There needs to be a more quantitative study and more thorough description of how to analyze certain behaviors to show a causal effect with energy supply demand. However, the fundamental premise of the proposal is questionable. Power plants are designed to follow the load. They will automatically ramp up or down depending on the demand for energy at any particular point in time. When an unusually large drop in demand occurs, such as when a transmission line fails, power plants will automatically trip. At any one point in time, the energy generated by power stations is exactly equal to the energy being used by consumers and, in some cases, by utility owned energy storage devices. This does not mean that more effective tools for predicting short (consumer side), and especially, long term, power needs would not be useful. However, it means that these tools will not allow utilities to reduce energy production in the manner suggested.
Aug 14, 2014
This is about Predictive Modeling/Forecasting Theory. It's important. It's effective and works. It's about power companies wanting to know when to conserve inputs to power generation. To quote from your project: "Shortly before the morning rush hour, using randomly-sampled mobility patterns and keyword searches from Twitter and other related sources, perhaps including a consumer application for the smartphone, statistical models predict that 20% fewer consumers will travel to work, and that instead they will stay at home. The power company finds with 90% confidence that they do not need to purchase power or turn on a plant that is normally required to supply downtown businesses, thereby saving a large amount of energy, CO2 emissions, and money. These savings can then be passed back to the consumer." Power companies should hire top-tier predictive modelers, gaming-scenario theorists?