Pro Membership

Pro Membership is a membership plan created by the author to maintain and update this tutorial. You can get more benefits and services, click to view details.

37. Stock Time Series Data Processing

37.1. Introduction

This challenge will use stock time series data to build a model through data processing and feature engineering to predict the rise and fall of stocks.

37.2. Key Points

• Time format conversion

• Resampling of time series data

• Data sorting

37.3. Challenge Introduction

In the time series processing experiment, we initially used the Google stock price change curve graph as an example for everyone. In this challenge, you need to apply the learned time series data processing methods to process the Google stock price change data.

37.4. Challenge Content

The challenge provides the stock trading data of Google from 2005 to 2018, which is aggregated into the data file GOOGL.csv. Download link:

wget -nc https://cdn.aibydoing.com/aibydoing/files/GOOGL.csv

A preview of the first 5 rows of the dataset is as follows:

Since there are trading days for stocks, the actual start and end dates of the data may not be from January 1st to December 31st.

	Date	Open	High	Low	Close	Adj Close	Volume
0	2005/1/3	98.798798	101.921921	97.827827	101.456459	101.456459	31656700
1	2005/1/4	100.800804	101.566566	96.836838	97.347351	97.347351	27484200
2	2005/1/5	96.821823	98.548546	96.211212	96.851852	96.851852	16456700
3	2005/1/6	97.637634	98.048050	93.953957	94.369370	94.369370	20753400
4	2005/1/7	95.415413	97.222221	94.484482	97.022018	97.022018	19306400

The data has 6 columns, namely: transaction date (Date), opening price (Open), highest price (High), lowest price (Low), closing price (Close), adjusted closing price (Adj Close), and trading volume (Volume).

{exercise-start}
:label: chapter04_04_1

The challenge requires performing time series analysis on this data file and sampling the data quarterly. Among them, the mean value is taken for all price-related data, and the total trading volume is taken. Finally, the DataFrame is sorted in descending order by trading volume and returned.

{exercise-end}

37.5. Challenge Requirements

1. The code needs to be saved in the Code folder and named google_stock.py.

2. The code in the following def quarter_volume() needs to be completed. At the end of the challenge, the required DataFrame needs to be returned.

3. A year can be divided into four quarters, each quarter lasting for 3 months. The first quarter: January - March; the second quarter: April - June; the third quarter: July - September; the fourth quarter: October - December. You need to learn how to perform quarterly sampling by reading the official documentation on your own.

4. When testing, run google_stock.py using python to avoid the situation of missing corresponding modules.

37.6. Sample Code

def quarter_volume():
    
    ### 补充代码 ###
    
    return df # 返回题目要求的 DataFrame

{solution-start} chapter04_04_1
:class: dropdown

import pandas as pd


def quarter_volume():
    df = pd.read_csv("GOOGL.csv", index_col=0)
    df.index = pd.to_datetime(df.index)
    df = df.resample('Q').agg({"Open": 'mean', "High": 'mean', "Low": 'mean',
                               "Close": 'mean', "Adj Close": 'mean', "Volume": 'sum'})
    df = df.sort_values(by='Volume', ascending=False)
    return df

{solution-end}

---

○ Sharethis article link to your social media, blog, forum, etc. More external links will increase the search engine ranking of this site.

If you find this content helpful, you can buy me a coffee

Previous

36. Time Series Data Analysis and Processing

Next

38. Time Series Data Modeling and Analysis