
Accountant and data analyst Also @peterdonaghy.bsky.social 🇺🇦
Belfast
Joined December 2008
- Tweets14.9K
- Following1.9K
- Followers9K
- Likes42.1K
Pinned Tweet
I've made a new website that tracks data over time to test claims about where the world is heading; things like "AI is going to disrupt the labour market" or "the current El Niño event will lead to higher food prices"
salmonofdata.com
Ok, so here's the *real* NI Assembly Election forecast: [your favourite party] will win all 90 seats. The First Minister will be [you] and the dFM will be [your best friend], you'll go on little adventures together. You'll have all the ice cream you want, they'll abolish bedtime,
My forecast model has SDLP as heavy favourites in Strangford. But - the model assumes the SDLP will outpoll SF in 1st preferences in Strangford in most model runs. In the scenarios where SF won more 1st prefs than SDLP, SF won a seat most of the time. So I could be wrong here.
Standing Chris Hazzard in Strangford isn't a reckless throw of the dice by Sinn Fein. It rarely makes election errors. This is a strategic, data-driven decision, & the party is confident it can win the first nationalist Stormont seat in the constituency. belfasttelegraph.co.uk/comme…
🚨 NEW: Andy Burnham has announced changes to car parking fines
- Drivers will have a 5-minute grace period to decide whether to park and 10 minutes to leave after a parking session ends
- A £20 maximum charge on appeal for minor mistakes, such as entering the wrong number plate or not displaying a permit
- Charges will also be cancelled on appeal for breakdowns, medical emergencies or broken ticket machines
- The £100 cap on charges will remain
- The rules will come into force as soon as possible and operators that break them will lose access to DVLA data
I ran SF with 2 candidates in Belfast South & Mid Down through 1000 runs of the forecast model and it didn't rate their chances of winning 2 very highly. The most likely outcome was no change.
Sinn Féin MP Pat Cullen to run in Assembly election as party bids to take two seats in South Belfast & Mid Down belfasttelegraph.co.uk/news/…
DAERA released new data on Lough Neagh algae blooms recently. Cranfield returned an estimated chlorophyll concentration of 1,377 µg/L in September, but the probe is only rated to 500µg/L, above which readings are unreliable. 3.6 Roentgen, not great not terrible etc.
From the Assembly election model - this was the estimated number of Executive ministries by each party and designation (assuming each party takes their entitlement and Justice isn't allocated by d'Hondt)
Belfast is wild because you can't get home by public transport after 9pm but the courts are open to adjudicate on parading disputes 24/7
I made a couple of changes to the Assembly Election forecast model. It has increased the estimated number of Unionist seats slightly, but Nationalist was the largest designation 68% of the time, Unionist 10% of the time, with a tie in 22% of the model runs.
At a designation level, this is how that would look in a historic context. An unchanged number of net Unionist seats, but Nationalist would be the biggest designation by taking 2 Unionist seats (UB & EL/D) and 2 Alliance seats (Strangford & LV).
This was the distribution of seat win probabilities by designation from the Assembly Election forecast.
The best case scenario for Nationalism was 45 seats - this happened in 0.5% of the 1000 model runs. It would look like this or something similar.
It's early days, but the slowdown in professional and business services hiring in the US and the decline in the software pay premium in Canada could be a sign that AI is beginning to have a measurable impact on jobs.
One thing I'm never ever doing again is an NI Local Election forecast model. Godspeed to all you public-spirited oddballs wanting to get involved with local government but I'm not getting involved. Thank you for your service though 🫡
A small change to the Assembly Election forecast model - it wasn't using a unionist transfer model for PUP votes. Fixed this and it had a small impact on EB, UB and EL/D, and increased overall unionist expected seats by 0.4.
The model assumes a unionist -> nationalist swing in Upper Bann. If this doesn't materialise an SF gain could be at the expense of Alliance.























