Forecast

Time series forecasting with Prophet

What is the Forecast?

The Forecast is a service that predicts future values based on your data history. It's like having a crystal ball for your business!

You provide your past data (sales, revenue, traffic, etc.) and the system automatically calculates how much you can expect in the coming days, weeks, or months.

Usage examples:

  • • Forecast next month's sales to plan inventory
  • • Estimate future revenue for financial planning
  • • Anticipate demand for products or services
  • • Project growth in visits to the website or app

Quick Start

  1. 1. Prepare your data in CSV or Excel format with two columns: date and value
  2. 2. Upload the file to the upload page
  3. 3. Set up the analysis parameters (frequency, periods, seasonality)
  4. 4. Please wait for processing (usually 2-5 minutes)
  5. 5. Visualize the generated graphs and metrics

How to organize your data

Organize your data in an Excel or CSV spreadsheet with two columns:

Column 1: Data

When did it happen? For example: 01/01/2023, 03/15/2024

Column 2: Value

How much was it? For example: 1500, 2300, 1890

Sales spreadsheet example:

data vendas
01/01/2024 1500
02/01/2024 1620
03/01/2024 1480
04/01/2024 1750

💡 Tip: You can use any names for the columns. It can be 'data_venda' and 'total', or 'dia' and 'receita'. The important thing is to have one column with dates and another with numbers.

Forecast Settings

Date column name

What is the name of the column that contains the dates in your spreadsheet?

Example:

data

Column name of values

What is the name of the column that contains the numbers you want to predict?

Example:

vendas

Frequency

How often do you record the data?

D Diary - every day has a record
W Weekly - once a week
M Monthly - once a month
Y Annual - once a year

How many periods to forecast

How far into the future do you want to predict?

Examples:

  • • If your data is monthly and you want to forecast 1 year: enter 12
  • • If your data is daily and you want to forecast 30 days: enter 30
  • • If your data is weekly and you want to forecast 8 weeks: enter 8

Recurrent patterns

Do your data have variations that repeat regularly?

Weekly standard

Example: lower sales on Sundays, higher on Fridays

Annual standard

Example: higher sales in December (Christmas), lower in January

Holidays

The system can take into account the impact of your country's holidays on the forecast.

Examples:

  • BR = Brazil
  • US = United States
  • PT = Portugal

Understanding the results

After processing your data, you will receive graphs that show not only the forecast but also the patterns found in your data.

Main Chart - Your Forecast

Black dots

These are the data you submitted. What has already happened in the past.

Blue line

It is the forecast. What the system believes will happen in the future.

Light blue area

Shows the margin of error. The actual value may vary within this area.

Other Charts - Patterns Found

Overall Trend

Shows whether your numbers are growing, declining, or remaining stable over time.

Example: your sales have grown 10% per month over the last 6 months

Standard of the Year

Shows in which months your values tend to be higher or lower.

Example: sales always go up in December and drop in January

Pattern of the Week

Shows on which days of the week your values tend to be higher or lower.

Example: sales are higher on Fridays and lower on Sundays

Impact of Holidays

Shows how holidays affect your numbers.

Example: sales drop 20% during Christmas and New Year's

⚠️ Important: The forecast is based on past patterns. If something changes significantly in the future (new competition, market shift, economic crisis), the results may be different.

Video Tutorials

Complete Tutorial - Forecast

How to Interpret Results

Need help? Contact us: contato@grabatus.com