Data for Climate Action Challenge / Campus Energy Analytics

Climate change potentially has raised the risk of flood, especially for coastal counties. Flooding can affect the living environment and threaten the success of crucial development schemes. In the UN Data for Climate Action challenge, we are focusing on the climate mitigation solution on road networks in Senegal. Road connections are vital for sustainable socioeconomic development, however under great vulnerability given the increasing incidence of flooding in recent years and the prevalence of unpaved roads. Our ultimate goal is to assess the effects of flooding events on road network connectivity, thus provide a budget-constrained optimization framework that can minimize the potential flood impacts.

The first stage includes a deep look into the flooding problem and a wide range of data collection and preparation. To build a model that can determine the flood risk of each road segment, we need to understand the relation between flood areas and the weather and topographical features. NASA (National Aeronautics and Space Administration) provides a daily global flood map in a high resolution of 250m*250m since 2013. We scraped the maps of Senegal from 2015 to 2017. The below picture shows a topographic map of Ziguinchor region and the flood areas (highlighted by yellow) change since 2015.

Weather data comes from NOAA (National Oceanic and Atmospheric Administration), which has archived daily weahter data of 12 major regions in Senegal since 1973. We obtained the geographic map of waterways and water areas in Senegal from OpenStreetMap, and elevation data from NASA SRTM (Shuttle Radar Topography Mission). Focusing on the Ziguinchor region of Senegal, we create 1km*1km grids as units for feature value extraction and labeling.

Composite Map with Flood, Waterways, Water Areas and Elevation.

The next step is to do basic statistics on grid samples, and build a preliminary model to assess the probability of flood based on historical weather data and topographic features.


Georgia Tech strives to be a leader in sustainability. The university is dedicated to promoting action and awareness of sustainable principles for the welfare of our environment. One of Tech’s planned initiatives to become a sustainable campus is reducing energy usage by 15 percent by 2020. To help Georgia Tech accomplish this, we are creating energy analytics to monitor how different factors affect building energy consumption and predict certain buildings that are using energy inefficiently.

This past week, we have been gathering weather data, energy usage data, and class schedules for an approximation of occupancy data to develop a model that will help predict how much energy a building will use at a certain point in time. Upon gathering energy data for the CULC building, our first step was to create some basic visualizations to get a sense of trends in energy consumption.

Our future steps involve finishing collecting data and starting to build a predictive model using all the features that can predict energy expenditure.