52052 - Machine Learning for Listing Generator
Budget: $250 – $750 CAD
We have a Laravel PHP based project for Short Term Office Rentals. We are looking to enhance our application by developing a feature that would assist our clients in creating Listing Descriptions for their properties.
Using OpenAI’s ChatGpt, we require a feature that will automatically generate the listing description for our clients.
The functionality will be similar to this solution: https://app.mappedby.com/listing-description-generator
Parameters for Consideration
The feature will generate a description using the following parameters:
• Listing City and Address
• Neighbourhood based on the address
• Category (listings are categorized based on industry. For example, an office may be for rent in a MEDICAL building)
• Amenities (randomly select three amenities to be included in the description
• Availability based on the Calendar and Availability settings. (For example: “Open 7 days a week, and available on a long term basis for weekly or monthly rentals”)
• Images
o Either randomly select a number of images to be used, from a specific folder or if possible;
o Search online for images that would fit the description and based on the parameters above, or;
o Digitally create a set of images based on the parameters, or;
o A combination of the options above
The intended use of this feature is when the Client is setting up their listing.
Part 2 - Auto Listing Generator
The purpose of this feature is to go populate the site with content that includes images, listing description, and basic contact and location information.
In essence, this feature should respond to the following instruction as an example:
• Search for Coffee Shops within 50 kilometers of this Zip/Postal Code.
• Check the system for any duplicate listings and eliminate from the results.
• Generate listings based on the results, and add images for each listing. If no image exists, use stock images from web or select folder
As a first generation feature, the system would first require the administrator to conduct the search parameters. Future versions should utilize machine learning functionality to actively search and create new listings on a continued basis.
Here’s a wireframe of how we think this section could be accomplished, but we are open to suggestions:
https://xd.adobe.com/view/4a73064a-80e2-4d7c-88de-ec035b85b5c0-1f20/
To ensure that you have fully read the scope, please ensure to include the word 'GUITAR' with your bid.
Using OpenAI’s ChatGpt, we require a feature that will automatically generate the listing description for our clients.
The functionality will be similar to this solution: https://app.mappedby.com/listing-description-generator
Parameters for Consideration
The feature will generate a description using the following parameters:
• Listing City and Address
• Neighbourhood based on the address
• Category (listings are categorized based on industry. For example, an office may be for rent in a MEDICAL building)
• Amenities (randomly select three amenities to be included in the description
• Availability based on the Calendar and Availability settings. (For example: “Open 7 days a week, and available on a long term basis for weekly or monthly rentals”)
• Images
o Either randomly select a number of images to be used, from a specific folder or if possible;
o Search online for images that would fit the description and based on the parameters above, or;
o Digitally create a set of images based on the parameters, or;
o A combination of the options above
The intended use of this feature is when the Client is setting up their listing.
Part 2 - Auto Listing Generator
The purpose of this feature is to go populate the site with content that includes images, listing description, and basic contact and location information.
In essence, this feature should respond to the following instruction as an example:
• Search for Coffee Shops within 50 kilometers of this Zip/Postal Code.
• Check the system for any duplicate listings and eliminate from the results.
• Generate listings based on the results, and add images for each listing. If no image exists, use stock images from web or select folder
As a first generation feature, the system would first require the administrator to conduct the search parameters. Future versions should utilize machine learning functionality to actively search and create new listings on a continued basis.
Here’s a wireframe of how we think this section could be accomplished, but we are open to suggestions:
https://xd.adobe.com/view/4a73064a-80e2-4d7c-88de-ec035b85b5c0-1f20/
To ensure that you have fully read the scope, please ensure to include the word 'GUITAR' with your bid.