Peter Dormaar's Avatar

Peter Dormaar

@peterdormaar.bsky.social

doing analytics projects on things I'm fascinated with.

240 Followers  |  426 Following  |  33 Posts  |  Joined: 12.12.2024  |  1.9252

Latest posts by peterdormaar.bsky.social on Bluesky

Post image

@coreyhogan.ca Hey Corey, I made a network diagram that shows the support of different demographics with the BCNDP. Should I make one for the federal Liberals? I think it would help with messaging and strategy, or at least providing context, for the Confed campaign.

07.04.2025 19:50 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 1    πŸ“Œ 0

I see why OpenAI is worried, but this needs to be taken into context. A lot of Deepseek's bright claims are likely a result of management accounting instead of an equal horse race between OpenAI and them

27.01.2025 17:25 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

Meanwhile, all of OpenAI's capx costs are allocated to building O1. The buildings, all the chips, cooling... they need to buy everything new. For example, cooling systems can cost between $5-$10 mil alone for a data centre, let alone the cost of the building and all the new chips.

27.01.2025 17:21 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 1    πŸ“Œ 0

The developers of DeepSeek are bitcoin miners who used their spare compute to develop this, so all of their capx costs are already sunk costs, and allocated to bitcoin mining. Their only costs allocated to this project may be R&D and electricity costs. Continued below...

27.01.2025 17:16 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 1    πŸ“Œ 0

On this DeepSeek news story: I really welcome the increased competition in the market. I want to mention, though, that their claimed development cost of $6mil, if we take them at their word, isn't an apple-to-apple comparison with Open-AI. I'll explain below:

27.01.2025 17:11 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 1    πŸ“Œ 0

I'm testing two different stock trading strategies with fake money. The first uses a categorical machine learning model, and the second has me copy/pasting stock financials into chat-gpt and getting a buy/sell recommendation. I'll update on how well they work.

04.01.2025 06:58 β€” πŸ‘ 3    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

I can also see this being helpful in identifying key demographics to design policy for and market to. Demographics that are grey but have many orange lines connected to them are likely prime targets to be persuaded to get on-side with the NDP. #BCNDP #cdnpoli #bcconservatives

02.01.2025 07:43 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
Post image

I can zoom in to see the demographics. Other than looking nice, this has helped me understand support levels better. For example, management, counterintuitively, were NDP supporters in another model. Here, I see they connect to many supportive demographics like having master's degrees.

02.01.2025 07:38 β€” πŸ‘ 2    πŸ” 0    πŸ’¬ 1    πŸ“Œ 0
Post image

This is a cloud of demographics and their relationships to eachother and the two main BC parties. Mostly for the purpose of beauty. Orange correlate with the NDP and blue correlate with Cons. Grey correlate with neither, and are essentially the battleground of policy and marketing. #BCNDP #CDNpoli

02.01.2025 07:29 β€” πŸ‘ 5    πŸ” 1    πŸ’¬ 1    πŸ“Œ 0
Post image

I got some pushback in Reddit about how unreadable the last chart was, so this is hopefully a simplified version. Orange demographics correlate with #BCNDP. Blue demographics correlate with #BCConservatives. Grey demographics correlate with no party. Size indicates median proportion. #BCpoli

31.12.2024 01:38 β€” πŸ‘ 4    πŸ” 1    πŸ’¬ 0    πŸ“Œ 0

Hey Kim, long time no see!

31.12.2024 00:49 β€” πŸ‘ 1    πŸ” 0    πŸ’¬ 1    πŸ“Œ 0

The size of the bubbles represents their mean proportion of ridings. For example, women are a large bubble because they make up around 50% of ridings.

30.12.2024 01:38 β€” πŸ‘ 2    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

This is a different method from my previous ones, so while many demographics are in the same spot as my previous tweets, some have moved. For example, real-estate workers have moved from the conservative camp to the NDP camp. Fist Nations have moved from NDP to unaligned.

30.12.2024 01:35 β€” πŸ‘ 5    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
Post image

This network diagram shows demographic groups that correlate with #BCNDP support and #BCconservative support. The grey ones don't have a high correlation to either, so they are for the picking for either party. Meanwhile they have to hold on to their bases. #CDNpoli #BCpoli

30.12.2024 01:31 β€” πŸ‘ 4    πŸ” 1    πŸ’¬ 2    πŸ“Œ 0
Post image

This made a lightbulb go off. "Womanness" has a very low positive correlation with #BCNDP support. The reason polls say women support them is important parts of their lives make them supportive. They're in jobs that support them. Men work jobs that overwhelmingly dislike the NDP. #CDNpoli #BCpoli

28.12.2024 07:01 β€” πŸ‘ 5    πŸ” 2    πŸ’¬ 0    πŸ“Œ 0

Single parents are also quite supportive. This makes sense because often they rely on social services. #BCNDP #BCpoli #CDNpoli

28.12.2024 06:21 β€” πŸ‘ 4    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

First Nations people are supportive, even though their effect in my model is quite small compared to others like employment industry. #BCNDP #BCpoli #CDNpoli

28.12.2024 06:15 β€” πŸ‘ 5    πŸ” 2    πŸ’¬ 0    πŸ“Œ 0

Marriage creates a small negative effect on #BCNDP support. Coefficient of -0.012904, R^2 value of 0.1449, P-value of 0.001 meaning it's statistically significant. #CDNpoli #NDP

28.12.2024 06:04 β€” πŸ‘ 3    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

I want to add that the individual variables aren't statistically significant in this one. I plan to make a more accurate model later, but it will take a lot of effort so I'm holding off.

28.12.2024 05:49 β€” πŸ‘ 3    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
Post image

While atheism has no effect on #BCNDP support, Christianity's effect is quite negative. This is still not very determinant of political support. #CDNpoli #BCPoli

28.12.2024 05:41 β€” πŸ‘ 5    πŸ” 1    πŸ’¬ 1    πŸ“Œ 0

The proportion of women in a riding positively correlates with #BCNDP support, although this factor alone is really bad at predicting support. So far Industry of employment is the most significant factor.

28.12.2024 02:03 β€” πŸ‘ 4    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

Whether BC voters have children or not is statistically insignificant in their voting decisions. This factor explains less than 1% of #BCNDP support. #BCpoli #ndp

28.12.2024 01:52 β€” πŸ‘ 2    πŸ” 1    πŸ’¬ 0    πŸ“Œ 0
Post image

Predictably, higher income groups supported the #BCNDP less than lower income groups. The R^2 of this is very low, meaning that income is a lot less determinant of support than other things, especially what industry people work in. #NDP #Canpoli

28.12.2024 01:42 β€” πŸ‘ 1    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
Post image

Youth support for #BCNDP is quite high, while their highest support is middle-aged adults. the support drops off with seniors. with an R^2 of 0.3, age isn't that big of a predictor, especially compared to career industry. #BCpoli #NDP

28.12.2024 00:15 β€” πŸ‘ 4    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
Post image

These new models are less predictive but are more accurate at explaining support levels of census groups. Here we can see levels of education's effect on #BCNDP support. Highschool certificates and tradespeople generally don't support them as much as other levels. #canpoli #BCpoli

27.12.2024 23:48 β€” πŸ‘ 2    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
Post image

I remade my model to be more explanatory. The industry someone works in is the best predictor of their support for the #BCNDP. this chart shows their support of them by percentage in riding in 2024 (intercept of 0.6594). #BCpoli #canpoli

27.12.2024 23:41 β€” πŸ‘ 2    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

Homeowners generally don't support the BCNDP as much as non-homeowners. Their presence increases support from resources sectors that are usually negative. Rural/resource industry homeowners may feel more stable with the incumbent gov't. #BCpoli #canpoli #BCNDP #NDP

27.12.2024 01:58 β€” πŸ‘ 1    πŸ” 1    πŸ’¬ 0    πŸ“Œ 0

I think this is an extremely important finding. Arts and entertainment grants that create jobs can persuade almost every kind of person in a riding to support the #BCNDP more. #Canpoli #NDP

27.12.2024 01:35 β€” πŸ‘ 1    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

The prevalence of arts an entertainment in ridings increases support, but also increases support from many other sectors, that don't usually support the BCNDP. Natural resources, agriculture, wholesale, and others all increase their support when there is more arts and culture in ridings

27.12.2024 01:32 β€” πŸ‘ 1    πŸ” 1    πŸ’¬ 1    πŸ“Œ 0

Even though BC Farsi communities generally don't support the NDP, their presence in communities increases the NDP support of healthcare and education workers. I think this could mean that Iranians value these sectors. Including these in messaging to Iranians may be interesting #BCNDP #Canpoli #NDP

27.12.2024 00:47 β€” πŸ‘ 1    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

@peterdormaar is following 17 prominent accounts