Plot BTC Market Cap Weekly EMAs

btc-mc-EMA-plot.py

import datetime as dt
import requests
import io
import pandas as pd
import matplotlib.pyplot as plt
from matplotlib.pyplot import figure
from matplotlib import ticker
from matplotlib.ticker import (MultipleLocator, FormatStrFormatter, AutoMinorLocator)
%matplotlib inline

# function to download CSV file from URL; convert to Pandas DataFrame
def get_csv_from_url(url):
    s = requests.get(url).content
    df = pd.read_csv(io.StringIO(s.decode('utf-8')))
    return df

if __name__ == '__main__':
    # data source: coinmetrics.io - all BTC community data (CSV)
    url = "https://coinmetrics.io/newdata/btc.csv"

    # download data
    print("Downloading data...")
    df = get_csv_from_url(url)

    # select columns
    df = df[['time', 'CapMrktCurUSD']]

    # rename columns
    df = df.rename(columns={"CapMrktCurUSD": "market cap"})

    # convert 'time' (object series) to timestamp format
    df['time'] = df['time'].astype('datetime64[ns]')

    # divide market cap by 1 billion
    df['market cap'] = df['market cap'] / 1000000000

    # rename for convenience
    x = df['time']
    y = df['market cap']

    # calculate Exponential Moving Averages (EMAs) [Periods: 5, 10, 20, 50, 100, 200]
    # enter (#-days) per period for EMAs, e.g. '7' = 1 week/period
    per = 7 # one week
    df['EMA_5p'] = df.iloc[:,1].ewm(span=(5*per),adjust=True).mean()
    df['EMA_10p'] = df.iloc[:,1].ewm(span=(10*per),adjust=True).mean()
    df['EMA_20p'] = df.iloc[:,1].ewm(span=(20*per),adjust=True).mean()
    df['EMA_50p'] = df.iloc[:,1].ewm(span=(50*per),adjust=True).mean()
    df['EMA_100p'] = df.iloc[:,1].ewm(span=(100*per),adjust=True).mean()
    df['EMA_200p'] = df.iloc[:,1].ewm(span=(200*per),adjust=True).mean()

    # rename for convenience
    y5 = df['EMA_5p']
    y20 = df['EMA_20p']
    y10 = df['EMA_10p']
    y50 = df['EMA_50p']
    y100 = df['EMA_100p']
    y200 = df['EMA_200p']

    # set starting and ending dates
    # start = len(df['time']) - (365*1+1) # one year ago
    # start = len(df['time']) - (365*2+1) # two years ago
    # start = len(df['time']) - (365*3+1) # three years ago
    start = 1308 # market cap >= $100M from this date onwards: '2012-08-03'
    end = len(x) - 1 # last date in series

    start_date = df['time'][start]
    end_date = df['time'][end]

    # get start/end market cap values (Billions USD)
    start_mc = y[start]
    end_mc = y[end - 1]

    # set ranges for plotting on chart
    x = x[start:end]
    y = y[start:end]
    y5 = y5[start:end]
    y10 = y10[start:end]
    y20 = y20[start:end]
    y50 = y50[start:end]
    y100 = y100[start:end]
    y200 = y200[start:end]

    # set figure size
    plt.figure(figsize=[24,12])
    ax = plt.subplot(1, 1, 1)

    # plot values in log scale
    ax.semilogy(x, y, label='BTC Market Cap', linewidth=3)
    ax.plot(x, y5, label='EMA 5')
    ax.plot(x, y10, label='EMA 10')
    ax.plot(x, y20, label='EMA 20')
    ax.plot(x, y50, label='EMA 50')
    ax.plot(x, y100, label='EMA 100')
    ax.plot(x, y200, label='EMA 200')

    # set y-scale limits: $100 million to $1.4 trillion
    ax.set_ylim(0.1, 1400)
    # ax.set_ylim(50, 300)

    # set x-scale limits:
    # start/end dates 10 days before/after specified `start_date`/`end_date`
    ax.set_xlim(start_date - dt.timedelta(days=10), end_date + dt.timedelta(days=10))
    # ax.set_xlim(start_date - dt.timedelta(days=5), end_date + dt.timedelta(days=5))

    # set Y-axis tick formats
    ax.yaxis.set_major_formatter(ticker.FormatStrFormatter('%.0f'))
    ax.yaxis.set_minor_formatter(ticker.FormatStrFormatter('%.0f'))
    ax.tick_params(which='major', width=3, length=8)
    ax.tick_params(which='minor', width=2, length=5)

    # for the minor ticks, show only every 100 billion mark
    ax.yaxis.set_minor_locator(MultipleLocator(200))

    plt.grid(True)
    plt.legend(loc=2, fontsize=16)
    plt.xlabel(f'Date Range: {start_date.strftime("%m/%d/%Y")} (\${start_mc:.2f}B) - {end_date.strftime("%m/%d/%Y")} (\${end_mc:.2f}B)', fontsize=16)
    plt.ylabel('Market Cap [Billions USD]', fontsize=16)
    plt.title("Bitcoin Market Cap vs. Weekly EMAs - Aug. 2012 to Present", fontsize=21)
    plt.show()

My matplotlib skills are underdeveloped. How about Altair?

$ pip install altair

import pandas as pd
import altair as alt
import requests

import datetime

# Quick and dirty helper for all those monetary values. We really should multiply all monetary
# values by 100 instead and represent it as an integer (and use /100 to print again).
pd.options.display.float_format = "{:6.2f}".format

alt.themes.enable("latimes")

resp = requests.get(
    "https://api.blockchain.info/charts/market-cap?timespan=all&scale=1&daysAverageString=7&format=json")
if resp.status_code != 200:
    raise RuntimeError(f"{resp.status_code}: {resp.content}")
data = resp.json()

df = pd.DataFrame.from_records(data["values"])
df = df.rename(columns={"x": "date", "y": "market cap"})
df["date"] = df["date"].apply(datetime.datetime.fromtimestamp)

# Set threshold of $1 billion USD or greater
df = df[df["market cap"] >= 1000000000]
df = df.dropna()

# Exponential Moving Average
df["50wk EMA"] = df.iloc[:,1].ewm(span=50,adjust=True).mean()
df["200wk EMA"] = df.iloc[:,1].ewm(span=200,adjust=True).mean()

alt.Chart(df).transform_fold(
    ["market cap", "50wk EMA", "200wk EMA"],
    as_=["leg", "USD"]
).mark_line().encode(
    x="date:T", y="USD:Q", color="leg:N"
).properties(
    title=f"{data['name']}"
)

Updated formatting.

Switch data provider to Coinmetrics.io, add EMA 10, update formatting.
https://coinmetrics.io/data-downloads/

2012 present chart, updated Jun 13, 2021 to reflect higher max market cap:

One year chart:

import datetime as dt
import requests
import io
import pandas as pd
import matplotlib.pyplot as plt
from matplotlib.pyplot import figure
from matplotlib import ticker
from matplotlib.ticker import (MultipleLocator, FormatStrFormatter, AutoMinorLocator)
%matplotlib inline

# Author: Adam Anderson, https://github.com/84adam
# Shout Out to: Benjamin Kircher, https://github.com/bkircher for help with formatting, data download

# function to download CSV file from URL; convert to Pandas DataFrame
def get_csv_from_url(url):
    s = requests.get(url).content
    df = pd.read_csv(io.StringIO(s.decode('utf-8')))
    return df

if __name__ == '__main__':
    # data source: coinmetrics.io - all BTC community data (CSV)
    url = "https://coinmetrics.io/newdata/btc.csv"

    # download data
    print("Downloading data...")
    df = get_csv_from_url(url)

    # select columns
    df = df[['date', 'CapMrktCurUSD']]

    # rename columns
    df = df.rename(columns={"CapMrktCurUSD": "market cap"})

    # convert 'date' (object series) to timestamp format
    df["date"] = df["date"].astype('datetime64[ns]')

    # divide market cap by 1 billion
    df['market cap'] = df['market cap'] / 1000000000

    # rename for convenience
    x = df['date']
    y = df['market cap']

    # calculate Exponential Moving Averages (EMAs) [Periods: 5, 10, 20, 50, 100, 200]
    # enter (#-days) per period for EMAs, e.g. '7' = 1 week/period
    per = 7 # one week
    df['EMA_5p'] = df.iloc[:,1].ewm(span=(5*per),adjust=True).mean()
    df['EMA_10p'] = df.iloc[:,1].ewm(span=(10*per),adjust=True).mean()
    df['EMA_20p'] = df.iloc[:,1].ewm(span=(20*per),adjust=True).mean()
    df['EMA_50p'] = df.iloc[:,1].ewm(span=(50*per),adjust=True).mean()
    df['EMA_100p'] = df.iloc[:,1].ewm(span=(100*per),adjust=True).mean()
    df['EMA_200p'] = df.iloc[:,1].ewm(span=(200*per),adjust=True).mean()

    # rename for convenience
    y5 = df['EMA_5p']
    y20 = df['EMA_20p']
    y10 = df['EMA_10p']
    y50 = df['EMA_50p']
    y100 = df['EMA_100p']
    y200 = df['EMA_200p']

    # set starting and ending dates
    start = len(df['date']) - (365*1+1) # one year ago
    # start = len(df['date']) - (365*2+1) # two years ago
    # start = len(df['date']) - (365*3+1) # three years ago
    # start = 1308 # market cap >= $100M from this date onwards: '2012-08-03'
    end = len(x) - 1 # last date in series

    start_date = df['date'][start]
    end_date = df['date'][end]

    # get start/end market cap values (Billions USD)
    start_mc = y[start]
    end_mc = y[end - 1]

    # set ranges for plotting on chart
    x = x[start:end]
    y = y[start:end]
    y5 = y5[start:end]
    y10 = y10[start:end]
    y20 = y20[start:end]
    y50 = y50[start:end]
    y100 = y100[start:end]
    y200 = y200[start:end]

    # set figure size
    plt.figure(figsize=[24,12])
    ax = plt.subplot(1, 1, 1)

    # plot values in log scale
    ax.semilogy(x, y, label='BTC Market Cap', linewidth=3)
    ax.plot(x, y5, label='EMA 5')
    ax.plot(x, y10, label='EMA 10')
    ax.plot(x, y20, label='EMA 20')
    ax.plot(x, y50, label='EMA 50')
    ax.plot(x, y100, label='EMA 100')
    ax.plot(x, y200, label='EMA 200')

    # set y-scale limits: $100 million to $1.2 trillion
    # ax.set_ylim(0.1, 550)
    ax.set_ylim(100, 1200)

    # set x-scale limits:
    # start/end dates 10 days before/after specified `start_date`/`end_date`
    # ax.set_xlim(start_date - dt.timedelta(days=10), end_date + dt.timedelta(days=10))
    ax.set_xlim(start_date - dt.timedelta(days=5), end_date + dt.timedelta(days=5))

    # set Y-axis tick formats
    ax.yaxis.set_major_formatter(ticker.FormatStrFormatter('%.0f'))
    ax.yaxis.set_minor_formatter(ticker.FormatStrFormatter('%.0f'))
    ax.tick_params(which='major', width=3, length=8)
    ax.tick_params(which='minor', width=2, length=5)

    # for the minor ticks, show only every 100 billion mark
    ax.yaxis.set_minor_locator(MultipleLocator(50))

    plt.grid(True)
    plt.legend(loc=2, fontsize=16)
    plt.xlabel(f'Date Range: {start_date.strftime("%m/%d/%Y")} (\${start_mc:.2f}B) - {end_date.strftime("%m/%d/%Y")} (\${end_mc:.2f}B)', fontsize=16)
    plt.ylabel('Market Cap [Billions USD]', fontsize=16)
    plt.title("Bitcoin Market Cap vs. Weekly EMAs - One Year", fontsize=21)
    plt.show()

updated to account for column name change: 'date' to 'time'