Weather Data Analysis

Inspiration

Throughout the years temperatures have risen, which also rapidly increased wildfire occurrences across the world. Most of them happen in places that are close to the equator, however even here in Slovakia we are not free of the risks of wildfires. In year 2010 Slovakia lost 5.38 kha of forest land due to wildfire. Main motivation behind our analysis is to raise awareness of the problem we are facing in the near future and also providing a building block from which we can learn, which places are most vulnerable to wildfires.

How we built it

This project consist of analysing 71 years of continuous precipitation measure- ments provided by IBL weather software solutions through API. Dataset that was used in this project is called ERA5-Land Historical Monthly Means by Hour of Day. Main goal is to use this data only from specific regions that have great masses of forest land and measure the concentration of lowest precipitation val- ues throughout the 71 years of data. Region with the highest percentage of drought days in the hottest months of the year is considered as more likely to develop wildfire.

Challenges we ran into

Greatest challenge could be the data retrieval. As programmer with little to no knowledge of climatology and weather sciences we had problem understanding what data we are even using, however through rigorous study of API and the weather now-casting and forecasting we were able to overcome that problem and land on an definite solution.

What we learned

We were able to understand more of how working with Big Data really is and that it can't be taken easily. Learning more about how technology and weather comes hand in hand with each other was a great learning jump, and thanks to that we now have wider range of view on data science in multiple fields. Furthermore, our skills in statistics were put to a great use and thoroughly tested, by which we have found our gaps in statistical analysis.

What's next for Weather Data Analysis

Our results are purely theoretical, although they bring some useful insights to the table, it is not useful to ordinary people. We have considered creating interactive web interface for searching through 71 years of precipitation data and highlighting regions endangered by wildfires more accurately and even visually to users.

More Results

To see the more results and conclusions to our work please check out the Analysis of Drought Periods in Regions of Žilina, Bánska Bystrica, Košice and Prešov pdf file in our GitHub repository.

Built With

• jupiter
• python