# Daily average Fee by year in USD

In this post we’ll see how it’s possible, with just a few lines in Python, to analyze how the average daily cost of fees changes over time, from one year to the next but also from one month to the next. As a reminder, we are analyzing the bitcoin blockchain.

Let’s start by importing the libraries.

``````import blocksci
import pandas as pd
import seaborn
import matplotlib.pyplot as plt
import matplotlib.ticker
import collections
import numpy as np
%matplotlib notebook``````

Instantiate the chain object:

``chain = blocksci.Blockchain("/BlockSci/config_file")``

We also use `converter `to express the value of fees in USD. Exchange rates are provided by CoinDesk.

``converter = blocksci.CurrencyConverter()``

We use the `fees_by_year()` function to define the year of interest and then be able to apply it to different years. This is the code:

``````def fees_by_year(year):

blocks = chain.range(year)
fees = blocks.fee / blocks.tx_count
times = blocks.time

df = pd.DataFrame({"Fee":fees}, index = times)
df = converter.satoshi_to_currency_df(df, chain)
ax = df.resample("d").mean().plot(legend=False)

return df, ax ``````

Now, we can apply the function by defining the year of our interest. Let’s try the years 2017, 2019, and 2020.

``df_2017,ax_2017 = fees_by_year("2017") ``
``df_2019,ax_2019 = fees_by_year("2019") ``
``df_2020,ax_2020 = fees_by_year("2020") ``

From the chart we can see that the average fee per transaction in 2019 has remained essentially stable. On the other hand, if we look at 2017, there has been a noticeable increase in the last few months. The reason for this is easy to guess, because as we know there was a significant growth in the popularity of bitcoin in 2017, this meant more purchases and an increase in competition for transaction validation. As we can see something similar happened in the last months of 2020.

Would you like to analyze the bitcoin blockchain using Python?

Apply to our Beta Tester Program!