Long story short – I got notified by my agency to vacate my current apartment in a month. Given the current COVID19 situation I am a bit skeptical around what rental availability looks like and the safest yet fastest way to job hunt.
This blog series aims to automate the gruelling effort of scrolling through listings and finding the next best place to stay. It is written according to SOAP guidelines and would be updated as I learn more about integration between these components.
The architecture sketch above is the overall solution view with each component functionality defined below. For simplicity I plan to use Azure Functions to take advantage of serverless capabilities & Google for location & schedule insights.
AU Rental Data
I remember seeing rent data for agreements signed in the last month with unit details. They can be used to create outliers in rent, post codes to focus/prioritize on and derive past trends.
Pending Question : What would be the format of the insights data to be used in DI?
Rental Listings
Get rental listings by postcode from aggregators like Domain.com.au or RealEstate.com.au for the last six months in the first extract run, followed by once per day. Looking for APIs or Python functions to extract this data instead of scraping information from the website.
Pending Questions :
- Same listing can be posted at multiple aggregators.How to manage duplicates ?
- When in the day do we want the extract to run?
- How do we manage for blank data?
- What are the failsafe measures ?
Decision Intelligence
This module would be responsible for making decisions on which retail listings qualify based on the criteria defined & market trneds. It would consist of the the following services that would filter the retail listings towards final list of qualifiers.
Unit : Compares retail listing properties against requirements such as rent per week, utilities etc.
Market : Checks local market trends for the particular post code to add value & weightage to the current retail listing
Location : Gather location insights about the property like distance from supermarkets, shopping malls etc. Most importantly it measures the total time taken to reach Sydney CBD across all mediums of transport
Pending Questions :
- What should be the grading/scoring method?
- Which feature should get more scores or have higher weightage?
- How do we manage data around images & disqualified listings?
Contact Queue
This module would manage the final interaction with the listings in the order of availability & alignment with critiera using the following services :
Avail Email : From past experiences, listing might not be updated with the latest availability. Upon qualification an email with draft questions would be sent out to the agency/owner to get further details.
Schedule : Automated system to query availability in my Google calendar and coordinate with property agency to set up time for inspection. Following travelling salesman problem, the schedules need to be optimized based on location of properties to be inspected
Pending Questions :
- What would be the user interface for me to check details of any qualified listing?
- How can we automated coordination via email for availability?
- Would it be wise to purchase SMS provider service for even more automation?