The Digital Observatory @ QUT

Notable from the party faithful too. Of course this allows them to attach blame with a figure who is jettisoned. Sky News, Bolt etc., going so far to say it was all Dutton, and LNP should lean more into culture wars (Credlin). Defies believe.

04.05.2025 00:19 — 👍 1    🔁 0    💬 0    📌 0

Let's return to our political discussion on Reddit data (roughly same volume as our NewsTalk dataset). Now let's consider support and oppose separately, and put LNP, ALP, Greens and All in the same plot. I include 'All' primarily to capture negative sentiment to all parties on key issues.

04.05.2025 00:16 — 👍 0    🔁 0    💬 0    📌 0

For people (rightly) skeptical of 'sentiment analysis'. This isn't it, this is full fat LLM processing of tens of thousands of comments, with each comment coded by 56-fine grained topics and the LLM tasked to assess likely political support or opposition based on a comment in a nested thread.

04.05.2025 00:11 — 👍 0    🔁 0    💬 0    📌 0

Political sentiment radar split by major topic grouping, with subplots for ALP and LNP net political sentiment by news website comment source.

I've applied our AI pipeline to process Australia news website comments (The Digital Observatory's NewsTalk product). I've broken it down by the top 8 news websites (for reader comments) with subplots for the ALP vs LNP.

04.05.2025 00:10 — 👍 1    🔁 0    💬 1    📌 0

Week 3 report might be a bit late since most of us are on leave next week, but here's top topic Reddit political party sentiment visualisation for week three of the election campaign anyway. Negative sentiment towards LNP re nuclear and renewables is well up this week. Greens support up re housing.

18.04.2025 05:27 — 👍 0    🔁 0    💬 0    📌 0

What do Reddit users discuss about Australia's Federal Election 2025? Updates on week 2 — Australian Digital Observatory

Our report on the 2nd week of the federal election campaign is now out www.digitalobservatory.net.au/blog/week-2-... (see the end of the written report for what's changed)

15.04.2025 23:37 — 👍 0    🔁 0    💬 0    📌 0

AI analysis of many thousands of comments by Australians on federal election topics, indicating political support/opposition by major parties. For more, see www.digitalobservatory.net.au/blog/what-do...

11.04.2025 03:50 — 👍 0    🔁 0    💬 0    📌 0

But by and large, there is a striking reduction in the complexity of language. I was motivated to look into this after seeing the the difference between Trump's earlier and contemporary speech in the source dataset. To show it in absolute terms:

06.03.2025 06:37 — 👍 1    🔁 0    💬 0    📌 0

In a sense yes. Lots of impromptu remarks being reported versus events with better planned speech, or speech written by others I expect. This was plotted from the 'interview' subset of the data to avoid the latter, I should have mentioned that.

06.03.2025 06:32 — 👍 1    🔁 0    💬 2    📌 0

This is lexical diversity plots of 40+ years of Donald Trump speech (~38 million words), organised by parts-of-speech. See bumps to adjective diversity when he took to inventing silly names for opponents. Data ripped from Factbase rollcall.com/factbase/.

06.03.2025 05:42 — 👍 3    🔁 1    💬 1    📌 0