# WaterGPT
**Repository Path**: theWinder/WaterGPT
## Basic Information
- **Project Name**: WaterGPT
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: main
- **Homepage**: None
- **GVP Project**: No
## Statistics
- **Stars**: 0
- **Forks**: 0
- **Created**: 2025-10-27
- **Last Updated**: 2025-10-27
## Categories & Tags
**Categories**: Uncategorized
**Tags**: None
## README
# WaterGPT
Examples of AI use in water operations.
> [!NOTE]
> The responses are not deterministic; you might need to adjust your prompts.
Example 1 - Missing data imputation
-----------
Upload the [flow excel file](data/flow_missing_data.xlsx)
> **User:** *plot the attached data*.
> **ChatGPT:** *Here is the plot of the flow data over time based on the provided dataset.*
> 
> **User:** fill in the missing flow data based on the daily pattern, plot again, and indicate the filled in data using red color
> **ChatGPT:** *Here is the updated plot showing the flow data over time, with missing values filled based on the daily pattern.*
> 
+ Validation of results in [Excel](data/example1_flow_original_missing_filled_data.xlsx).
Example 2 - Importing, processing, and visualizing asset data
-----------
Upload the [pipes and junctions files](data)
> **User:** *Open the zip files and examine the files*.
> **ChatGPT:** *the contents of the zip files are as follows: ...These files are components of shapefiles typically used in geographic information systems (GIS) for mapping and spatial analysis. Would you like to explore the data within these files further? .*
> **User:** *Plot the water pipes and junctions*.
> **ChatGPT:** *Here is the plot showing the water pipes and junctions. The blue lines represent the pipes, and the red dots represent the junctions*.
> 
Upload the [meters file](data)
> **User:** *This is a water meters layer. Load the file and add the meters to the plot*.
> **ChatGPT:** *Here is the updated plot, now including the water meters. The green "x" markers represent the meters, alongside the blue pipes and red junctions*.
> 
> **User:** *I want to associate each meter with the closest junction. Plot a gray line connecting the meters with their associated junction*.
> **ChatGPT:** *The plot now includes gray lines connecting each water meter to its closest junction*.
> 
> **User:** *I want to associate each meter with its closest pipe. Add a line perpendicular to the pipe*.
> **ChatGPT:** *The plot now includes gray lines connecting each water meter to its closest pipe with a perpendicular connection*.
> 
> **User:** *Save a file with the closest junction and pipe for each meter*.
> **ChatGPT:** *The associations between meters, closest junctions, and closest pipes have been saved to a CSV file. You can download it using the link below:*
[meter_associations.csv](data/meter_associations.csv)
+ Steps to validate the results in [ArcGIS](data/example2_GIS_write-up.docx).
Example 3 - Analyzing AMI data
-----------
Upload the [consumers data](data)
> **User:** *Load and describe the data*.
> **ChatGPT:** *The data from the sheet "Demand May" consists of the following columns: ...*
> **User:** *Sum over different meter numbers and plot the total demand over time*.
> **ChatGPT:** *The plot above shows the total demand over time, summing the demand values across different meter numbers...*.
> 
> **User:** *Show the mean demand profile for 24 hours for different days of the week*.
> **ChatGPT:** *The plot illustrates the mean demand profile for 24 hours across different days of the week. Each line represents a day of the week, showing the average demand at each hour. ...*.
> 
> **User:** *This data is from Ein Mahil is an Arab local council in Israel, it has a population of 13,931, the majority of which are Muslims. Can you explain the early peak on Friday?*.
> **ChatGPT:** *The early peak in demand on Friday can be attributed to the cultural and religious practices of the Muslim community in Ein Mahil. For Muslims, Friday is considered a holy day, known as Jumu'ah. It involves special congregational prayers (Jumu'ah prayers) that take place around midday. ...*.
> **User:** *Make boxplot for different hours of the day, remove outlier*.
> **ChatGPT:** *The box plot above shows the demand distribution for each hour of the day, with outliers removed. ...*.
> 
> **User:** *Repeat and remove negative values*.
> **ChatGPT:** *The box plot now shows the demand distribution for each hour of the day, with outliers and negative values removed ...*.
> 
> **User:** *Make histogram of average daily demands of the users, remove outliers and negative values*.
> **ChatGPT:** *The histogram now shows the distribution of average daily demands for users, with both outliers and negative values removed ...*.
> 
> **User:** *Make boxplot of mean daily demand for top 10 users, remove negative values and outliers*.
> **ChatGPT:** *The box plot displays the mean daily demand for the top 10 users, with outliers and negative values removed ...*.
> 
+ Validation of results in [Excel](https://utexas.box.com/shared/static/ya56kmg581fqz8oek60yljlcf0yhgyfa.xlsx).