My main input into this data sprint was in the project’s initial data wrangling and background research. We had two datasets available to us on PosgreSQL, one that described all wind turbines in the U.S and another from the U.S. Energy Information Administration that had data on power operators. I started my wrangling by looking at the individual tables, and I found that it would be easiest to understand the data as one table. The dataset about wind turbines had a good amount of information, but it lacked the supplemental detail on plant operators that could provide information on power projects in the country as a whole, including how expansive they were, who was managing them, and how many projects an operator was managing. As such, I decided to use a left Join to merge the operator data that was relevant to wind turbines to the wind turbine data set.

The resulting dataset, while useful, needed to be manipulated with more sophistication than PosgreSQL would have allowed. I decided to export the joined table to a CSV file and then upload it to Jupyter Notebook. Using the Python Pandas library, I was able to look at our 63,000 rows of EIA data with more clarity. To clean the data, me and my teammates filtered our dataset to check for any null values and where they might be. I took the lead in determining what to do with these values and found that there were over 15,000 rows of entirely null data - these were all dropped. Many others represented turbines missing data on their plant’s physical unit label, but seeing as we were focused on wind power exclusively I decided that these null values should be replaced with the dataset’s unit for gaseous energy sources: mcf. Other null values had to do with a turbine’s geography, which we filled in by checking the state of a turbine and matching it to its geographical region.

The dataset being otherwise clean, we were able to export this updated table to a CSV file to begin visualizations. Meanwhile, I turned my attention to the background research necessary to contextualize our findings. I focused on investigating the science behind wind power and the current wind turbine market.

Understanding how wind power is generated required that I focus my search to more technical and scientific sources. Much of the information gathered here came from either the energy.gov website or research-based pages. Through these I was able to provide detail to our team’s work on the types of wind-turbine configurations used to generate electricity, as well as the physics and constraints behind them. The latter was especially important, as it lent a deeper understanding to the importance of turbine height, wind speed, and blade length and how all three reflected varying power outputs by different turbine manufacturers and power plants.

Then, to understand the renewable energy market as a whole, I turned mostly to the eia.gov website to provide useful insight and data files on changes in U.S energy markets over time. By finding and investigating this data, I was able to present a market incentive for investment by demonstrating the growth of renewable energy sources and the multiple times higher generation potential of wind power.