It would be nice if we could see the last 10 years.

Here is total employment for "Computer Systems Design and Related Services" going back all the way to 1987. This is the best category I can find that includes data over a decade+ time frame.

https://fred.stlouisfed.org/series/IPUMN54151W200000000

This category is defined by the BLS as "establishments primarily engaged in providing expertise in the field of information technologies through one or more of the following activities: (1) writing, modifying, testing, and supporting software to meet the needs of a particular customer; (2) planning and designing computer systems that integrate computer hardware, software, and communication technologies; (3) on-site management and operation of clients' computer systems and/or data processing facilities; and (4) other professional and technical computer related advice and services."

Yeah, using 2020 as a baseline for most industries is fine, but in the case of software the anomalous job market really started by 2014.

OK chart axis is biased see here: https://www.bls.gov/opub/ted/2024/job-openings-rate-at-4-9-p...

This is job openings in general, OP's is software development.

I really wish you could see this in the context of the prior decade when the economy was relatively stable.

It takes a bit of digging to find the definition of the index they’re using.

From their GitHub page: “The data in this repository are the percentage change in seasonally-adjusted job postings since February 1, 2020, using a seven-day trailing average.”

if you scroll down, looks like same shape as charts below it for other jobs that can be done remotely: banking, marketing, research, etc.

what am i missing?

Those graphs finish higher than the initial dip. The software one does not. But yes, the same overall shape.

Clear shape of a bubble? What are your thoughts?

The graph is a bit misleading, at least on mobile. The y-axis starts at 60. The data is also limited.

Edit: It's base 100 so my critique is unfair.

The graph is a bit odd on desktop too. The baseline is Feb 1, 2020, which equals 100 units on the y-axis. Given this normalization, I am less concerned about the y-axis starting at 60.

However, the graph would be much more informative if there were 10 or 20 years of data. It's unclear to me why Feb 1, 2020 is a good baseline to use.

This is an index with base 100.

I missed that, thanks for pointing it out.

The y index is very misleading, from people who should know better

Yes, but the peak is 3x the index prior to the bubble and now.

Theory:

- job titles change

- indeed search demand constant

- other platforms

- 2022 height of bubble

I doubt all the tracked job titles are in the example, but they do aggregate job titles to the category of 'Software Development':

https://github.com/hiring-lab/job_postings_tracker/blob/mast...

And they do aggregate from other sources:

https://www.hiringlab.org/indeed-data-faq-2/

It seems every major job posting platform gets overran with spam and eventually candidates move to a new platform every ~5 years.

First it was CareerBuilder, then Monster, then Dice, and now Indeed is having its day in the sun too. I wonder if the emergence of GenAI in the past 2 years have accelerated the spamification of Indeed and thus its demise. I’ve posted there and within minutes get hundreds of AI-generated “candidates” with practically the same resume/CV.

Can we see more data?

Here's the graph for all job postings in the US: https://fred.stlouisfed.org/series/IHLIDXUS Note all job postings were up in 2022 so if the point was about software in particular it should probably be normalised.

OK clearly biased by the selection: https://www.bls.gov/opub/ted/2024/job-openings-rate-at-4-9-p...

Hadn't seen this one. Thanks for sharing

Yeah I think the signal is much smaller than it looks. UK saw a similar overshoot following COVID: https://www.ons.gov.uk/employmentandlabourmarket/peopleinwor...

"data in this graph are copyrighted." No, no they are not. Data can't be copyrighted, at least in the USA. The format of the days can be copyrighted. But not the data itself.