The candidate for the job must understand and use Machine Learning, it needs to use something like Azure or AWS to learn from the data.
I have a construction / building materials company (hardware, tools, etc) and I need to keep an up to date database of the products we have for sale along with their prices and description. The products are sorted into categories. For instance product 'hammer' belongs in category 'tools' with the 'type' of hammer, a 'description' and the 'price' of the hammer.
There's multiple stores that send a CSV file each month with their hardware and the price they are selling it for. Sometimes its completely new items that need to be inserted, other times it's just a price that needs updating.
I need to be able to upload multiple .csv files and have the rows of data automatically (as best as possible) be sorted and stored in the MySQL database.
There's a base set of data which contains two tables, 'category' table and then 'items' table.
I would like the system to take the .csv files and apply some kind of machine learning rules to the data, compare it to our base set of data, it would then take the new .csv rows and insert them into the 'items' table and then it would tag each row with the associated 'category' that is the closest / best match (based off the learning from the base data set).
This is a small project with similar logic to what I'm after: [login to view URL]
The end goal is to have a simple interface where I can upload all the CSV files as they arrive and then the system will automatically update the main MySQL database by inserting new items (and tagging them to their predicted category) and update existing items by inserting their new price (compare the name of the item to existing items). I need an automated updating system so I don't have to manually check and update thousands of rows. The CSV columns do not always line up to the database columns, so the importer may need a step where the user maps the columns to the database - that is OK - if it can be ML guessed / predicted that would be even more of a bonus.
I have the skills to create the uploaders / importer / table mapping side of things in PHP, the main thing I need is the ML component.
I dont mind what language or service the main process for the ML / sorting procces is written in, so long as it can run on a digitalocean droplet and have a web interface for uploading the .csv files (super basic uploader), then store to MySQL, the rest is whatever you feel is best. ie: ML platform can be AWS or Azure, blah, whichever you think is best.
Bonus if this can be achieved using mostly PHP and JS as I have a better understanding of these languages.
Can provide creds to a DO droplet for dev work etc, will also provide data set for learning, example CSVs and MySQL structure etc to the awarded freelancer.