Green_Energy_Locations

💭 Idea:

It is unclear where the optimal locations for wind and solar energy are. For example this could be a problem for homeowners who plan to build a solar system.

💡 Potential solution:

Analyze solar and wind data. Also consider infrastructure and legal requirements to create a map with optimal locations.

1. Idea and Goal:

Analyse wind and solar data
Look at legal requirements as well as meteorological data (wind speeds, sunlight hours, etc.) and consisting infrastructure (high voltage lines)
Show these locations a map (Highlight the best locations with a colour-ranking)
Optional: Calculate the amount energy that can be produced at a specific location
Optional: Calculate the break even-point

→ Goal: Recommend ideal locations for solar or wind parks

2. Description

In the following, we are going to describe our understanding of the business question or problem, analytical question and/or technical problem associated with the use case.

What are the challenges that need to be mastered?

2.1. Setup Github structure:

Issues & Cards on Kanban-Board
Create an environment where all group members can work together and stay up-to-date

2.2. DataFrame:

Find a suitable database, to download relevant meteorological data for wind and solar
Create a Dataframe, which contains relevant input variables for the machine learning model

2.2. Solar-Model:

Which input-variables for the model:
- Solar: hours of sun
- How many hours of sunlight is needed for a solar power plant?
  - Therefore, we chose the hours of sunshine and the clearness of the sky
What is the output-variable and how is it defined?
- Is the location suitable for photovoltaic?
- If the location is better than 60% of the other stations, it is defined as Suitable
- Train different Machine-learning models and fit them with the help of Hyperparametertuning and a cross-validation
- We chose the random forest because it performed best
- Notebook for Data Preparation: code/solarmodel_means.ipynb

2.3. Wind-Model:

Which input-variables for the model:
- average wind speed
What is the output-variable and how is it defined?
- Is the location suitable for a windmill?
- If the mean windspeed of the location is higher than 60% of the other stations, it is defined as Suitable
Train different Machine-learning models and fit them with the help of Hyperparametertuning and a cross-validation
- How do we get the most accurate model?
- Training the model with the standard deviation input variables results in the highest accuracy
- -> compare accuracy of different models:
- logistic_regression_train shows the highest accuracy score we still chose to use the random_forest_train because log_reg led to a large number of negatives
- Notebook for Data Preparation: code/windmodel_deviations.ipynb

2.4. Map:

Setup Map:
- Plot a map of Germany
- Plot each location with latitude and longitude on the map
- Whether a location is suitable for a solar-plant or windmill, it is plotted as a green dot -Whether a location is not suitable for a solar-plant or windmill, it is plotted as a red dot
Compare the plots with maps that show the actual windmills and solar plants in Germany