In this repository I did exploratory data analysis of foreign exchange rates. This dataset was generated on the Federal Reserve's Download Data Program. Some changes were made in the dataset, such as header simplifications and inversions of base currency. For example, Fed provides Australian Dollar, Euro, New Zeland Dollar and United Kingdom Pound based in their units (not in dollar).
Link to kaggle: https://www.kaggle.com/brunotly/foreign-exchange-rates-per-dollar-20002019
Github link to repository: https://github.com/MadoDoctor/EDA-Project-Foreign-exchange
In this exploratory data analysis we worked with packages:
- NumPy;
- Pandas;
- Matplotlib;
- Seaborn;
- CSV;
- Datetime.
We removed the first column as it is repeating the index, then taked general info as shape and columns type.
We divided the "Time Serie" column into days, months, years and days of the week to find out on which day of the week the data is most absent:
We calculated on which day of the week the data are most absent, dropped all rows with no data and checked in which day of the week was the maximum and minimum value of Euro and the maximum and minimum values for each year.
Firstly we dropped all rows with no data. Then changed the column names to make them understandable. We taked mean value for each currency:
Mean value for each currency was to all years, we can take mean value for every year if we want.
We moved all column names to one column and their values to another, by using matplotlib.pyplot we plotted lineplot:
Then we sorted currencies in descending order of their values. By sorting, we can take information like the most expensive or the cheapest currency in our data.