Context

I prepared this dataset for a project on rainfall forecasting. Met Éireann is the Irish Meteorological Service and a scientific organisation that undertakes research in numerous fields such as Numerical Weather Prediction and Climate Modelling. Their short-term predictions work very well. I was curious to see how Machine Learning and Deep Learning models would handle these types of tasks. Care to join me?

Description

This dataset contains:

• Folder with Individual CSV files for 24 Met Éireann weather stations in Ireland capable to record hourly weather data (the start date of individual time series depends on when the particular station was opened, the end date is 2022-02-01);
• Aggregated hourly weather data from 24 stations in Ireland for the period of time from 2007-12-31 to 2022-02-01 (with station names and locations added);
• The list of stations.

Variables measured by stations (available variables may vary depending on the station):

• date: Date and Time of observation
• ind: Encoded Rainfall Indicators (see KeyHourly.txt for details)
• rain: Precipitation Amount, mm
• ind.1: Encoded Temperature Indicators (see KeyHourly.txt for details)
• temp: Air Temperature, °C
• ind.2: Encoded Wet Bulb Indicators (see KeyHourly.txt for details)
• wetb: Wet Bulb Air Temperature, °C
• dewpt: Dew Point Air Temperature, °C
• vappr: Vapour Pressure, hPa
• rhum: Relative Humidity, %
• msl: Mean Sea Level Pressure, hPa
• ind.3: Encoded Wind Speed Indicators (see KeyHourly.txt for details)
• wdsp: Mean Hourly Wind Speed, knot
• ind.4: Encoded Wind Direction Indicators (see KeyHourly.txt for details)
• wddir: Predominant Hourly wind Direction, degree
• ww: Synop Code Present Weather (see KeyHourly.txt for details)
• w: Synop Code Past Weather (see KeyHourly.txt for details)
• sun: Sunshine duration, hours
• vis: Visibility, m
• clht: Cloud Ceiling Height (if none value is 999), 100s of feet
• clamt: Cloud Amount, okta

>"Wind direction is usually reported in cardinal (or compass) direction, or in degrees. Consequently, a wind blowing from the north has a wind direction referred to as 0° (360°); a wind blowing from the east has a wind direction referred to as 90°, etc." >Wikipedia page for "Wind direction"

Table: Common Cardinal (or compass) direction vs degrees

Information on the stations:

• county: County the station is losated in
• st_id: Station number
• st_name: Station name
• st_height: Station Height, m
• st_lat: Station Latitude, sexagesimal degrees (degrees, minutes, and seconds - DMS notation)
• st_long: Station Longitude, sexagesimal degrees (degrees, minutes, and seconds - DMS notation)

Latitude and longitude are presented in sexagesimal degrees (degrees, minutes, and seconds - DMS notation). To convert them into decimal degrees (DD) which are used in GIS and GPS apply the following formula: DD = D + M/60 + S/3600. More details can be found here.

Acknowledgements

Data were obtained from the Met Éireann website.

Copyright statement: Copyright Met Éireann Source: www.met.ie Licence Statement: This data is published under a Creative Commons Attribution 4.0 International (CC BY 4.0). Disclaimer: Met Éireann does not accept any liability whatsoever for any error or omission in the data, their availability, or for any loss or damage arising from their use.

Hourly weather data for 24 stations were downloaded and aggregated into one dataframe. Station names and locations were added.

Photo by Nils Nedel on Unsplash was used as a Banner image.

Inspiration

For EDA and Data Visualization:

1. What are the most prominent seasonal weather patterns in Ireland?
2. How does the weather conditions affect city life?
   • Pedestrian footfall
   • Bikeshare sevices
   • Road accidents
   • Taxi

For ML and Neural Networks modelling:

1. Can you predict the probability of rain using weather data obtained from a single station in the previous 24, 36 or 48 hours?
2. How does the addition of data recorded by neighbouring stations affect the accuracy of the model?

Articles for ideas:

1. Streamflow and rainfall forecasting by two long short-term memory-based models
2. Short-Term Rainfall Forecasting Using Multi-Layer Perceptron
3. Monthly Rainfall Forecasting Using One-Dimensional Deep Convolutional Neural Network
4. A hybrid support vector regression–firefly model for monthly rainfall forecasting