Every month since Feb 2020. Real data only: grey land = no comparable data.
FEB 2020
EUROPE ZOOM
Map: Natural Earth 1:50m (public domain), Robinson projection, cropped 56°S to 76°N, Antarctica left out. One pixel is roughly 300 km. Countries with data get at least 2 pixels (3 in the Europe zoom), so very small ones are drawn larger than they are. Dashed frame = Europe zoom.
One equal tile per country, roughly where it sits on the globe (own layout, from Natural Earth points). Only countries with data light up.
MORE LAYOFFSMORE HIRINGNO DATANO SCORE, HAS OPEN ROLES
FEWMANY OPEN TECH ROLESNONE TRACKED
IS YOUR COMPANY HIRING? PUBLIC JOB BOARDS · DAILY SNAPSHOTS
Pick the month you started and your country. We compare it with the latest official figure on the same series. No figure for your month means no card number: we do not guess.
Pick a role and a place. Role trends exist only where Indeed publishes them (US, UK, Canada, Australia, Germany, France; Data & AI not for France); cities get their all-jobs trend plus today's open roles.
Open-role counts from public job boards, for editorial use. Credits & licences
LOWEST/HIGHEST SINCE… COMPUTED FROM EACH SERIES
STILL HIRING TOP 5
GUESS THE YEAR 5 REAL MONTHS
The map below is a real month from the since-2000 view. Which year is it?
What we count: open roles on each company's public Greenhouse, Lever or Ashby job board, read once a day through those tools' public job-board APIs (no sign-in, no scraping). "Tech roles" use the same title rules as LIVE OPEN ROLES below; duplicates (same title and locations) count once. "All open roles" is every posting on the board.
Function split sorts tech roles by job title (software, data & AI, IT/cloud/security, product & design, hardware); some titles will be misfiled. Places are the posting's listed locations matched to GeoNames cities; a role in 3 cities counts in each.
Posted in the last 30 days uses the date the job board reports: Greenhouse "first published", Lever "created", Ashby "published". Boards can repost old roles, so read it as a rough freshness signal.
The stamp:
History so far:
Not tracked? Most large employers (Google, Amazon, Microsoft, Apple, Meta, banks) use other hiring systems with no public API, so they are not here. For ~540 well-known tech companies we looked for a public board; if none was found we say so. We never estimate numbers for a company we do not track. Layoff counts per company are not shown: we only hold Layoffs.fyi's published totals, not its per-company table.
THE MAP
Each lit tile gets a score from −1 (red: layoffs win) to +1 (green: hiring wins), measured against that country's own pre-pandemic normal.
Hiring, monthly (10 countries): Indeed job postings index (Feb 2020 = 100), averaged per month. Score part H = log2(index ÷ 100), capped at ±1. So 200 (postings doubled) = +1, 100 = 0, 50 (halved) = −1. Software development postings for US, UK, Canada, Australia, Germany and France. Hiring Lab publishes no sector data for Ireland, Spain, Italy and the Netherlands, so those use all-job postings (marked ALL JOBS).
Hiring, quarterly (everywhere else we can): Indeed publishes nothing for the rest of Europe or New Zealand. For 26 more European countries we use the Eurostat job vacancy rate in the ICT sector (NACE section K: telecoms, programming, IT consulting, data centres, information services). For New Zealand we use MBIE's IT industry job ads index. Both are quarterly and not seasonally adjusted, so each quarter is compared with the same quarter of 2019: H = log2(value ÷ same quarter 2019), capped at ±1. A quarter's value is shown for all three of its months. Eurostat rounds rates to one decimal, so countries with very low rates (Bulgaria, Greece, Croatia, Slovakia) move in big steps.
Layoffs: US: BLS JOLTS layoffs and discharges rate in the Information sector (monthly). UK: ONS redundancy rate (monthly, all industries). 28 European countries: Eurostat's quarterly job-loss rate, the share of people in work last quarter who are unemployed this quarter (seasonally adjusted, all industries; no tech-only official series exists). Score part L = log2(rate ÷ 2019 average), capped at ±1. Values Eurostat flags as low reliability are dropped. Germany has no 2019 baseline in that series, so it stays hiring-only.
Tile score = (H − L) ÷ 2 where both exist, otherwise H alone (marked HIRING ONLY).
Official data comes out later than job postings. For the newest months we reuse the latest published figure and say so: monthly layoffs at most 3 months (LAYOFFS AS OF …), quarterly series at most 4 months past the quarter (LATEST PUBLISHED / LAYOFFS AS OF Q…). Gaps in the middle of a series stay gaps. Nothing is estimated or filled in. While playing, colours fade between real monthly values for smoothness; every labelled month is a real value.
Mixing sources: postings, vacancy rates and ad counts are different measures, so compare each country's direction against its own past, not the raw numbers against each other. The source used is printed next to every country in the list.
Worldwide counter: tech employees laid off per quarter from Layoffs.fyi. It counts by company HQ, not by where the jobs were, so it is shown as one global number and does not colour countries.
Blue-grey tiles (latest month only): no index score, but the live open-roles snapshot found tech jobs there. Switch to LIVE OPEN ROLES to see them.
ChatGPT marker (30 Nov 2022) is shown as a question only. These data cannot say whether AI caused any change in hiring; rates, the end of the 2021 boom and other factors moved at the same time.
Grey pixels (dark tiles in the tile view): no reusable official or open tech hiring series for that country (we checked Indeed, FRED, Eurostat and national statistics offices).
TILES, GAUGE, JOIN CARD, QUIZ, PIXEL DROP
Pixel map (default view, v8): a pixel-art world map rasterised from Natural Earth 1:50m country shapes (public domain), Robinson projection, with a Europe zoom. Grey pixels are land with no comparable data. Countries with data get at least 2 pixels (3 in the Europe zoom): Singapore, Luxembourg, Malta and others are drawn bigger than they really are so you can see and tap them. Overseas territories far from the mainland (French Guiana, Svalbard, the Canaries and so on) are drawn as grey land and not coloured with the national figure.
Tile map (TILES button or ?view=tiles): every country is one equal square so small countries (Malta, Luxembourg, Cyprus) are as visible as big ones. Tiles sit roughly where each country is: our own layout, made by placing Natural Earth country label points (public domain) on a stretched grid and moving each to the nearest free square. Neighbours are approximate. Tile size says nothing about country size.
Is it worse than 2008? For countries whose official series reach back far enough, we compare the latest figure with the worst point in two windows on the same series: the dot-com bust (Mar 2001 to Dec 2003) and the 2008 crash (Jul 2008 to Dec 2010), plus the 2019 average. Monthly, seasonally adjusted series use the worst single month. Quarterly series that are not seasonally adjusted use the worst value for the same quarter as the latest figure. YES means the latest figure is worse than the 2008-crash point. When the worst dot-com point falls after 2001 (the US openings low is Apr 2003), it is labelled the dot-com aftermath with its date. A window is only used if the series starts within 6 months of it; no series, no gauge. Monthly US figures jump around, so the 3-month average is shown next to the latest value.
When did you join tech? Your start month on the same official series as the gauge (tech where it exists, otherwise all industries, labelled), against the latest figure. "Peak", "lowest" and "top 10%" are ranks within that series since it started. Quarterly unadjusted series compare your quarter with the latest figure for the same quarter. Where Indeed publishes postings (from Feb 2020) we add that line too. No figure for your month: the card says so.
Guess the year shows the since-2000 tile map for 5 real months, each revealed with figures from BLS JOLTS and ONS for that month. Choices are other years in the data.
Pixel drop: one pixel per 100 workers named in WARN notices, per month for the 12 months to the selected month, rounded to the nearest 100 per month (so a month under 50 workers shows no pixel). California and Texas only.
Look: fonts Press Start 2P (The Press Start 2P Project Authors) and VT323 (The VT323 Project Authors), both under the SIL Open Font License 1.1, served from this site (licence, licence). Sound is off by default and only plays after you tap; it uses the ambient audio setting so a phone's silent switch mutes it.
SINCE 2000 LONG VIEW · SEPARATE SCORE
Why separate: Indeed postings only start in Feb 2020, so the long view uses only official statistics that reach back before 2010. It is a different score from the main map; the two are not spliced together.
Baseline: each series is compared with its own long-run average up to 2019 (from Dec 2000, or from the series start if later). Quarterly series that are not seasonally adjusted are compared with the average of the same quarter across all years up to 2019. H = log2(hiring value ÷ its average), L = log2(layoffs value ÷ its average), each capped at ±1; score = (H − L) ÷ 2, or one side alone where only one exists (marked HIRING ONLY / LAYOFFS ONLY).
United States, from Dec 2000: BLS JOLTS Information sector job openings rate (hiring) and layoffs and discharges rate (layoffs), monthly, seasonally adjusted. The counter shows Information-sector layoffs and discharges per month (thousands).
United Kingdom: ONS vacancies in information and communication (JP9P, from May 2001) and the ONS redundancy rate (all industries, from Dec 2000 here).
Europe: Eurostat ICT job vacancy rate (quarterly; from 2001 to 2012 depending on country) and the Eurostat job-loss rate (all industries). Countries whose data starts after 2016 are left dark in this view.
Australia, from Nov 2009: ABS job vacancies, Information Media and Telecommunications (quarterly, original). ABS suspended the survey from Aug 2008 to Aug 2009, so there is no 2008 crash for Australia. New Zealand, from Dec 2010: MBIE IT job ads index.
Japan, from 2002, layoffs side only: Statistics Bureau of Japan Labour Force Survey, unemployed people whose reason for looking for work is dismissal or business circumstances (all industries). We found no current information and communications hiring series in the e-Stat API. Japan's dismissals have fallen far below their 2002–2019 average, so its tile sits near +1: read it as "few job losses", not "tech is booming".
Era labels (dot-com bust, global financial crisis, COVID) use the NBER US recession dates and are only a guide for other countries.
Then vs now figures are calendar-year averages of the published values; a year is shown only when the series covers all of it. For quarterly, unadjusted series we compare the same quarter instead.
Latest values may be reused for at most 3 months (monthly) or 4 months past the quarter (quarterly), marked LATEST PUBLISHED. Nothing is estimated or filled in.
Data from the U.S. Bureau of Labor Statistics (bls.gov). BLS.gov cannot vouch for the data or analyses derived from these data after the data have been retrieved from BLS.gov.
This service uses the API function of e-Stat, but the contents of the service are not guaranteed by the government. このサービスは、政府統計総合窓口(e-Stat)のAPI機能を使用していますが、サービスの内容は国によって保証されたものではありません。
LIVE OPEN ROLES SNAPSHOT · NOT COMPARABLE WITH THE SCORE
What it is: a count of open tech roles on the public job boards of a fixed, published list of companies (~300 companies, list kept with the code), read once a day from the public Greenhouse, Lever and Ashby job-board APIs. No sign-in, no scraping of job sites.
Tech role = title match: software, data, ML/AI, IT, security, infrastructure, product management and design. Sales/solutions engineers, recruiters and physical-engineering roles (mechanical, propulsion, manufacturing…) are excluded. Title rules are in the code and will misclassify some roles.
Where: each posting's listed locations are matched offline to GeoNames cities (pop. ≥ 15,000) and countries. A role listed in 3 cities counts once in each. "Remote" with no country and region-only roles (EMEA, Worldwide) are not placed. Duplicate postings (same company, title and locations) count once.
Coverage bias: this is not the whole market. The list leans to US and European venture-backed and big tech firms that use these three job-board tools; most large employers (Google, Amazon, Microsoft, Apple, banks, Indian IT services) use other systems and are missing, and a few big hirers (e.g. defence-tech) dominate some cities. Read it as "where these companies are hiring today", not as country size or a trend against 2019. It is a snapshot (); daily snapshots are stored so a trend can build.
US LAYOFF NOTICES (WARN)
What it is: notices employers must file before mass layoffs or closures (WARN Act). Only California (EDD reports) and Texas (Texas Workforce Commission open data) are included: other states either publish no reusable file or block downloads. Counts are workers named in the notice; the date is the date the state received the notice, not the layoff date.
Tech employers = company name matches our list of tech companies or contains a clear tech keyword (software, semiconductor, cybersecurity…). Staffing contractors, hotels, warehouses etc. working for tech firms are excluded. Some tech layoffs are filed under subsidiary names we do not recognise, so tech counts are a floor.
Where: the notice's city, matched to GeoNames (pop. ≥ 15,000); otherwise the largest city in the notice's county. Dots show the 12 months up to the selected month.
Gaps: Texas data published to ; California notices before July 2019 are incomplete.
YOUR JOB + YOUR CITY, LOWEST/HIGHEST SINCE, STILL HIRING
Roles are offered only where data exists. Indeed Hiring Lab publishes job-postings trends by category for six countries only (US, UK, Canada, Australia, Germany, France): Software Development, Data & Analytics (not France) and IT Infrastructure, Operations & Support. "Any tech role" uses the country score. Product & design and hardware have no public trend series anywhere, so they get open-role counts only.
No city-level role trends exist. A city card's big number is that city's all-jobs Indeed postings vs Feb 2020 (593 US, 58 UK, 45 Canadian metros). The role trend on the card is the national figure and is labelled "across" the country. "Worse/better than X% of N cities" compares the same month across all metros Indeed publishes for that country.
Open roles per role come from the same job-board snapshot, sorted into roles by job title (rules in scripts/roles.py; some titles will be misfiled). Ranks are among places with at least one such role.
US places also show the state's BLS JOLTS openings and layoffs rates (all industries; no industry split exists by state) and WARN tech layoff notices for California and Texas cities.
"Lowest/highest since…" is computed from each series: we look back from the latest value through the unbroken run of data and find the last time it was this low (or high). Only claims spanning 12+ months are shown. Quarterly unadjusted series compare the same quarter only. If the value is a record for the whole series we say "the lowest in … data" with the series start, never earlier. Indeed months are averages of daily values, and the current month is partial.
Still hiring: countries ranked by the latest score (Bulgaria, Greece, Croatia and Slovakia have very low ICT vacancy rates rounded to one decimal, so small moves look big); cities ranked by all-jobs Indeed postings vs Feb 2020 among metros whose main city has 250,000+ people (GeoNames), so one-employer towns do not top the list; and places with the most open tech roles in the snapshot.
CHECK YOUR CITY PROTOTYPE
Indeed all-job postings for 593 US metros, US state job openings and layoffs rates since 2000 (BLS JOLTS, all industries; type a state name too), 58 UK cities and 45 Canadian metros (Feb 2020 = 100; all jobs, Indeed publishes no city-level tech series), plus today's open tech roles at tracked companies for ~300 cities worldwide, plus WARN layoff notices for California and Texas cities.