[FIXED] Matplotlib plotting descending graph with Bitcoin price data

Issue

I am trying to plot the Bitcoin historical data on a Matplotlib graph using the this (2010-2024 CSV file) dataset.

When I graph it using Matplotlib, it gives the attached graph.

I have tried the same exact code with a different dataset (using the Solana crypto price dataset, it has the same exact CSV keys, just different values), and its graph is attached.

Here is my code:

import matplotlib.pyplot as mp
import pandas as pd
import numpy as np
import sklearn as sk
import random
from sklearn.linear_model import Perceptron

csv = pd.read_csv("./bitcoin-historical-trading-data.csv")
model = Perceptron()

csv['Date'] = pd.to_datetime(csv['Date'])
csv['Date'] = (csv['Date'].astype('int64') / 10**9).astype('int64') * 1000
csv = csv[["Date", "Price"]]

date = np.array(csv["Date"])
price = np.array(csv["Price"])

mp.plot(date, price)
mp.show()

Please let me know if you have any idea why this is happening.

Solution

When you see a Matplotlib plot with that many ticks, the first thing you should consider is that your data might not be numerical, see the Matplotlib FAQ.

To verify, print out csv.dtypes and I'm guessing the "Price" column will be of object type. Looking at the links you provided, the Bitcoin data is in the thousands while the Solana is not. Without an account I cannot download the data, but if the saved data has those commas, Pandas will read them as strings (and make them objects). Remove those commas and cast the data to a float.

csv["Price"] = csv["Price"].str.replace(",", "").astype(float)

Answered By - jared