Django Prophet: Building time series analysis applications from beginner to advanced-Python Tutorial-php.cn

Django Prophet: Building time series analysis applications from beginner to advanced

PHPz

Release： 2023-09-26 13:36:16

Original

2093 people have browsed it

Django Prophet: 从入门到高级，打造时间序列分析应用程序

Django Prophet: From entry to advanced, building a time series analysis application requires specific code examples

Time series analysis is an important statistical analysis method. It is used to study the changing trends, periodicity, seasonality and outliers of time series data. With the development of data science and machine learning, time series analysis has become increasingly important in areas such as forecasting and studying market trends and economic indicators.

Django Prophet is a Python-based time series analysis tool that combines statistical methods and machine learning technology to provide easy-to-use and highly customizable time series forecasting functions. This article will introduce how to use Django Prophet to build a time series analysis application and provide specific code examples.

Installing Django Prophet

First, we need to install Django Prophet. Open a terminal or command prompt and run the following command:

pip install django-prophet

Copy after login

Create a Django project

Next, we need to create a Django project. Run the following command in the command line:

django-admin startproject timeseries_app
cd timeseries_app

Copy after login

Create a Django application

Run the following command in the timeseries_app directory to create a Django application named timeseries:

python manage.py startapp timeseries

Copy after login

Then add 'timeseries' in the INSTALLED_APPS list in the settings.py file as follows:

INSTALLED_APPS = [
    ...
    'timeseries',
    ...
]

Copy after login

Create a time series model

In Create a models.py file in the timeseries directory and define a model class named TimeSeries, as shown below:

from django.db import models

class TimeSeries(models.Model):
    timestamp = models.DateTimeField()
    value = models.FloatField()

    def __str__(self):
        return self.timestamp.strftime('%Y-%m-%d %H:%M:%S')

Copy after login

This model class contains two fields: timestamp and value, which respectively represent the timestamp and the corresponding value.

Data preparation

In Django projects, we usually use the Django management background to manage data. Write the following code in the admin.py file in the timeseries directory to be able to add and manage TimeSeries model data in the management background:

from django.contrib import admin
from timeseries.models import TimeSeries

admin.site.register(TimeSeries)

Copy after login

Data upload

Start Django develops the server and logs in to the management background to upload time series data. Enter the following URL in the browser:

http://localhost:8000/admin

Copy after login

Then log in with the administrator account, click the "Time series" link, and click the "ADD" button in the upper right corner of the page to add a time series object.

Time Series Analysis

Next, we will write code in the view function to analyze and predict the uploaded time series data. Open the timeseries/views.py file and add the following code:

from django.shortcuts import render
from timeseries.models import TimeSeries

def analyze_time_series(request):
    time_series = TimeSeries.objects.all()

    # 将时间序列数据整理为Prophet所需的格式
    data = []
    for ts in time_series:
        data.append({'ds': ts.timestamp, 'y': ts.value})

    # 使用Django Prophet进行时间序列分析和预测
    from prophet import Prophet
    model = Prophet()
    model.fit(data)
    future = model.make_future_dataframe(periods=365)
    forecast = model.predict(future)

    # 将分析结果传递到模板中进行展示
    context = {
        'time_series': time_series,
        'forecast': forecast,
    }

    return render(request, 'analyze_time_series.html', context)

Copy after login

In the above code, we first get all the time series data from the database and organize it into the format required by Django Prophet. Then create a Prophet instance to fit and predict the data. Finally, the analysis results are passed to the template.

Template design

Create a template file named analyze_time_series.html to display the analysis results of time series. Write the following HTML code:

<!DOCTYPE html>
<html>
<head>
    <title>Analyze Time Series</title>
</head>
<body>
    <h1>Time Series Data</h1>
    <ul>
        {% for ts in time_series %}
            <li>{{ ts }}</li>
        {% empty %}
            <li>No time series data available.</li>
        {% endfor %}
    </ul>

    <h1>Forecast</h1>
    <table>
        <tr>
            <th>Timestamp</th>
            <th>Predicted Value</th>
            <th>Lower Bound</th>
            <th>Upper Bound</th>
        </tr>
        {% for row in forecast.iterrows %}
            <tr>
                <td>{{ row.ds }}</td>
                <td>{{ row.yhat }}</td>
                <td>{{ row.yhat_lower }}</td>
                <td>{{ row.yhat_upper }}</td>
            </tr>
        {% endfor %}
    </table>
</body>
</html>

Copy after login

In the above template, we use the template engine provided by Django to display time series data and prediction results.

URL configuration

The last step is to configure the URL routing so that we can access the analysis page through the browser. Add the following code to the urls.py file in the timeseries_app directory:

from django.contrib import admin
from django.urls import path
from timeseries.views import analyze_time_series

urlpatterns = [
    path('admin/', admin.site.urls),
    path('analyze/', analyze_time_series),
]

Copy after login

Run the application

You can now run the Django application and view the time series analysis results. Run the following command in the command line:

python manage.py runserver

Copy after login

Then enter the following URL in the browser:

http://localhost:8000/analyze

Copy after login

You will see the page of time series data and forecast results.

The above is all about using Django Prophet to build a time series analysis application from entry to advanced. Hopefully this article will provide you with practical code examples about time series analysis and Django Prophet, and help you further explore the world of time series analysis.

The above is the detailed content of Django Prophet: Building time series analysis applications from beginner to advanced. For more information, please follow other related articles on the PHP Chinese website!