Real Estate Technology
Market News & Data
General Info
Real Estate Strategies
Landlording & Rental Properties
Real Estate Professionals
Financial, Tax, & Legal
Real Estate Classifieds
Reviews & Feedback
Software Engineer Real Estate Investors
Any other software engineers become Real Estate Investors? Would love to hear about ways you've leveraged your skills for automations, integrations etc.
Hey Nicholas,
Not exactly a software engineer, but I am a Python / Data guy.
1. I pulled county home sales in my target area, and called an API that converts address to Lat and Long. From this, I made a map that showed which areas had the highest increase in home prices, to show the "path of progress", ie: which areas are on the edges of the appreciating areas that might appreciate next. It point out a new neighborhood that wasn't on my radar before, but ultimately two deals fell through there and I bought somewhere else.
2. I built a real estate calculator that works probabilistically. IE: instead of saying maintenance is $200 / month, it uses a distribution, most months it's $0, while some months its $2000. It gives a range of outcomes for each investment: IE: in 10 years the 80% confidence interval for ROI is 2-14%. At some point I'd like to build this into a website, but html is such a pain the butt.
As for automations, I just use Mint, and then export the data and use excel to automatically categorize things - not a very Software Engineer type solution, but it makes my accounting pretty quick.
Quote from @Stefan D.:
Hey Nicholas,
Not exactly a software engineer, but I am a Python / Data guy.
1. I pulled county home sales in my target area, and called an API that converts address to Lat and Long. From this, I made a map that showed which areas had the highest increase in home prices, to show the "path of progress", ie: which areas are on the edges of the appreciating areas that might appreciate next. It point out a new neighborhood that wasn't on my radar before, but ultimately two deals fell through there and I bought somewhere else.
2. I built a real estate calculator that works probabilistically. IE: instead of saying maintenance is $200 / month, it uses a distribution, most months it's $0, while some months its $2000. It gives a range of outcomes for each investment: IE: in 10 years the 80% confidence interval for ROI is 2-14%. At some point I'd like to build this into a website, but html is such a pain the butt.
As for automations, I just use Mint, and then export the data and use excel to automatically categorize things - not a very Software Engineer type solution, but it makes my accounting pretty quick.
Hey Stephan! Wow very cool Love the path of progress concept! Definitely something I will need to try out! Does your probabilistic model spread the maintenance costs out randomly for you?
The main thing I've worked on/planning to work on more is a payment portal for renters that also allows them to put in maintenance requests as well. I'd like to integrate a few different APIs to automate my discovering of new potential properties.
Hi Nicolas and Stephan,
I started my career as a Software Engineer. I've build a app that crawls foreclosure data, and county records like Stefan. I also crawl some of the listing sites, market reporting companies like NAR, and economic and demographic data in the North East.
So far I can get land records, rental estimates, owner data, median income, crime rates, comps. It can do a pretty good deal calculation (cashflow, roi). Its' pretty good and showing value add on properties that are on the market
Biggest challenges are that zillow is terrible with their data. There are a lot of properties that they provide data for that are really rental data. So a duplex will show up 3 times and look like 1 multifamily and 2 single units for only $1000 each.
As a small developer I can't afford the $1000+/month for API feeds from CoreLogic, AttomData, and Zillow. Once this app is marketable I could convert to published APIs.
Quote from @Nicholas Spinazze:
Any other software engineers become Real Estate Investors? Would love to hear about ways you've leveraged your skills for automations, integrations etc.
Yup. I am one of them. Happy to chat as I have done some on the automation front to simplify my life.
Yes, it spreads it out and runs a large number of simulations. You can choose the "skew" of the distribution (ie: is maintenance a normal distribution around $200/mo or is it a distribution with a long right tail), which is definitely something only super nerdy number people would care about - so maybe if I marketed this tool I would try to keep it a little more simple and hide some things under advanced settings.
Have you looked at the existing tools for that yet? I have to imagine there's some good ones out there already.
Nice work getting the crawling to work. I got blocked on trying to scape all the main real estate websites. If those main API Feeds can charge so much I bet there's room for you to make some money undercutting them.
Hi ,
I am a software engineer by profession , but I am only starting out in real estate. Currently I am house hacking , living in SFH and renting out the ADU. I would like to invest more in RE using my software skills which is what I am exploring. I have tried to develop some web scraping applications but as others have mentioned web scraping is not allowed by most RE websites. I would love to help out anyone in building software if you have ideas you want to try out.
automation is great! data is great!
i like to reflect on the fact that all of the resale RE data aggregators get their data for free from the county assessors roll... hilarious, no?
these are public records - our records - and are available to us for free, or a nominal production fee.
GIS records are usually free as well, eliminating the need for geocoding and including not only parcel boundaries, but a long list of parcel and building attributes, such as grade and condition, sq footage, yr of const, etc. very sweet, and open source GIS software aptly handles the task.
re scrapes - you can easily scrape any site that is humanly accessible, even if robots.txt and other means try to keep you out - look at scrapy, beautiful soup etc along with some vpn proxy farms. or you can hire this out.
and why scrape when the county will email you an 80mb csv? for free! lol!
converting monthly capex to a probabilistic distribution seems like it is the same thing, unless i am missing something... 0 0 0 0 0 0 0 0 0 0 0 2000 is the same as 200 x 12, right? maybe above my pay grade, but this seems true...
i just hired out a 100k scrape for a govt site that limited gets to 1k requests - got the data in about 20 hours, he used 93 proxies. hired this out bc the public records request was taking some time to complete, and the price was very cheap.
i am not sure that showing recent high appreciation paths is solid, and i speak from some experience. one way to test this is to model YOY data for this and see if it tracks, which it likely won't as this is more of a random distribution. it sounds like you have the chops to do this - if you model it, let us know the result for a few years of data...
i use these same ideas and test them out - one was to model the churn rate of different areas, the rate of sale each year. i found that these are varied as well YOY, although there is some tracking, ie high net worth areas turn at about 3% annually, and lower socioeconomic areas turn 6-8% per year - past that, this analysis was not too helpful for finding sellers.
at the end of the day, we want to find sellers.
i believe that there is no better source of data than the county - they are the originators, and they are required to provide us copies - daily, if we want!
i can extract any asset class right now from a dataset of 629k parcel records... duplex, apartments, 4plex, vacant commercial, mobile home, etc... and do almost anything i want with that data, like export it to a Google Earth KMZ file (parcel boundaries) with all attributes listed in the description, and hyperlinks to the county tax lookup, etc.
a great way to qualify parcels for development potential, and more.
all the best!
Free data, courtesy of the county....
Quote from @Nicholas Spinazze:
Any other software engineers become Real Estate Investors? Would love to hear about ways you've leveraged your skills for automations, integrations etc.
Hello Nicholas,
Fifteen years ago, I started by acquiring both public and paid data to identify potential properties for a specific tenant pool segment. When using such data, you are trying to predict where people will want to rent and how much they are willing to pay. However, the correlation between my predictions and reality was poor.
So I turned the process around, starting with the type of tenants I wanted. I identified a narrow tenant segment that had a high concentration of what I call "good tenants." I define a good tenant as someone who:
- Has stable employment in a market segment that is very likely to be stable or improve over time
- Pays all the rent on schedule
- Takes care of the property
- Does not cause problems with neighbors
- Does not engage in illegal activities while on the property
- Stays for many years
Once I identified a specific segment, I then derived what I call a property profile. The segment-specific property profile elements are:
- Location - Locations where significant percentages of the target segment live today.
- Property type - What type of properties are they renting today? Condo, high rise, multi-family, single family?
- Rent range - What the segment is willing and able to pay. Usually, this is about 30% of their gross monthly household income.
- Configuration - Two bedrooms, three-car garage, large back yard, single-story, two stories?
- Wants - Physical features the segment values. For example, granite kitchen counters, bars on the first-floor window, or a three-car garage.
With a property profile, I narrowed the search to a relatively small number of properties. I then wrote the software (Python) to automate the evaluation of each property. This enabled me to quickly eliminate properties that aren’t likely to perform and focus on the few with potential.
Over the years, I've accumulated 10+ years of behavioral data down to individual streets and subdivisions for the areas we normally find properties. Tenant segment-specific historical data is key to making good investment decisions.
What are our results over the last 15 years using data science?
- Delivered over 470 investment properties, >$120,000,000 delivered.
- 2013 through 2021 appreciation averaged >15% and rent growth averaged >8%. In 2022, rents increased by 4.5% but prices remained flat.
- More than 90% of clients buy more than one property from us, and over 80% buy more than two. This tells me that our clients are happy with their results.
- Our average tenant stays over five years.
- We've had six evictions in the last 15 years (over 1,200 tenants).
How dependable has our clients’ income been?
- 2008 crash - Zero decline in rent and zero vacancies.
- COVID - Almost no impact
- Eviction moratorium - Almost no impact
I put our overall investment methodology into a guide. Email me if you would like a free copy.
-
Real Estate Agent NV (#S.0067069)
Quote from @Stefan D.:
Hey Nicholas,
Not exactly a software engineer, but I am a Python / Data guy.
1. I pulled county home sales in my target area, and called an API that converts address to Lat and Long. From this, I made a map that showed which areas had the highest increase in home prices, to show the "path of progress", ie: which areas are on the edges of the appreciating areas that might appreciate next. It point out a new neighborhood that wasn't on my radar before, but ultimately two deals fell through there and I bought somewhere else.
2. I built a real estate calculator that works probabilistically. IE: instead of saying maintenance is $200 / month, it uses a distribution, most months it's $0, while some months its $2000. It gives a range of outcomes for each investment: IE: in 10 years the 80% confidence interval for ROI is 2-14%. At some point I'd like to build this into a website, but html is such a pain the butt.
As for automations, I just use Mint, and then export the data and use excel to automatically categorize things - not a very Software Engineer type solution, but it makes my accounting pretty quick.
To try and help your accounting, I do bookkeeping on the side and I have found Wave to be a decent free software. you can link your accounts to it and "teach" it to code for you so maybe it would help you to take less time to do your accounting and have it formatted right away for you.
-
Public Accountant
@Cliff Benner Thanks for the recommendation! I'll check out Wave