Overview

WinnipegCast is an application for visualizing the historical forecasts of the weather conditions in Winnipeg, Manitoba.

The application is split into three main parts:

1. An Angular client interface that queries a RESTful API backend server by date and displays the data that returns from the server call.
2. A Node/Express server which connects to a PostgreSQL database and services the client-side requests.
3. A PostgreSQL database that stores the weather forecasts.

Dependencies

• Node.js comes prepackaged with a Node Package Manager (NPM) client which is used to install required modules and dependencies for the application.

• Express is a lightweight Node.js web application framework for building web and mobile applications.

• Sequelize is an ORM for SQL databases.

• Dotenv is a module that loads environment variables from a .env file into a local or cloud environment.

• Multer is a middleware for handling forms and multi-part data.

• Nodemon is a tool that automatically restarts a Node.js application when file changes are detected.

• Pg is a PostgreSQL client for Node.js and Pg-hstore is a companion module for serializing and deserializing JSON data to hstore format.

• Angular is the framework used for designing the user interface of the application.

Installation

• Clone the source code for the application at https://github.com/Emmanuel289/WinCast.git.

• Navigate to the backend/ folder within the folder tree and execute npm install && npm start to install the dependencies for the backend and start the backend service.

• Navigate to the frontend/ folder and execute npm install && ng serve --open to install the dependencies for the web application and start the frontend service.

• (Optional) Start a postgres client using the psql startup shell and connect to the database after entering the required configuration parameters. You can also manage existing connections using the pgAdmin management console.

Querying the Weather Database for Forecasts

• The weekly_forecast.js located inside the backend/models/ folder contains the schema for defining and composing a weekly forecast into a table for storage in the database.
• The routes.rest file contains a list of standard HTTP methods for querying the local server which forwards the requests to the database. For example, the following request adds a new record to the database:

Content-Type: application/json

{
   "date_time_local": 2024,

   "temperatures": "High 25"

}

• Postman can be used to upload a csv file of datasets to create a new bulk record in the database or to update an existing record. The screenshot below shows the process of uploading a file named "weatherstats_winnipeg_forecast_daily.csv" into the database using Postman:

Postbird is a PostgreSQL client that can be used to connect to the PostgreSQL instance and view the uploaded dataset as shown below:

Design Considerations and Future Iterations:

• PostgreSQL was the choice of the database engine because the weather forecasts are relational and we might want to persist their storage, instead of having in-memory storage. Future iterations will introduce caching on the client side for more recent or frequently accessed data and archival of older records in a cloud datastore like BigQuery or S3.
• The user interface of the application is a skeleton in its present form and does not do much. It will be updated with a service that connects with the backend to display weather data to a user.