Scenario 1 · Income
Do people in richer areas tweet more happily about money?
The team asked whether conversations about pay, housing, debt and work sound different in high- and low-income parts of Victoria. Here the 2022 tweets that mention money are pooled from suburbs into the ABS SA2 areas that carry official income figures, so both can be compared on one map.
- Victorian SA2s analysed
- 420
- 37 removed as outliers
- Median income range
- $29.0k–$62.0k
- Merbein to Sydenham
- Income-related tweets in Victoria
- 27,999
- of 93,055 Australia-wide, Feb-Jul 2022
- Income vs sentiment (Pearson r)
- 0.14
- 167 SA2s · p = 0.07
Explore
Income on the map, sentiment on the chart
Every dot is an SA2 that had at least one tweet in the chosen category; bigger dots carry more tweets. Raise the threshold to drop areas whose average rests on a handful of posts. The correlation is recomputed in your browser with the same formulas the build script ran in scipy.
- Pearson r
- 0.14
- p = 0.07
- Spearman ρ
- 0.10
- p = 0.19
- R²
- 0.019
- of variance
- Regions
- 167
- in the fit
Computed in your browser; identical to the scipy values stored by the build script.
With at least 1 income-related tweet per SA2, income explains 1.9% of the variation in average sentiment: a weak positive association. Wealthier areas do not tweet noticeably happier, which matches the team's 2023 conclusion.
Select an area on the map or a dot in the chart (or search above) to see its income and tweet sentiment.
2026 upgrade · uncertainty
How sure can we be?
The correlations above are point estimates. Here the same comparison carries a 95% interval, a heteroskedasticity-robust slope and a check for leftover spatial structure, first at the 2023 threshold and then counting only SA2s with at least 30 tweets, where an average can bear some weight.
Income tweets, 2023 threshold
SA2s kept by the IQR rule with at least one income tweet: the comparison the 2023 page reported.
- Areas
- 167
- ≥ 1 tweets
- Spearman ρ
- 0.10
- 95% CI -0.06 to 0.26
- OLS slope
- 0.29
- HC3 95% CI -0.05 to 0.62
- p (HC3)
- 0.09
- robust, two-sided
- R²
- 0.019
- SE 0.160 classical vs 0.171 HC3
- Residual Moran's I
- -0.021
- p = 0.37 (permutation)
The interval for Spearman's rho runs from -0.06 to 0.26 and includes zero: the data are consistent with no association.
All tweets, SA2s with ≥ 30 tweets
The overall tone of SA2s whose averages rest on enough tweets to mean something.
- Areas
- 147
- ≥ 30 tweets
- Spearman ρ
- 0.20
- 95% CI 0.04 to 0.36
- OLS slope
- 0.13
- HC3 95% CI 0.03 to 0.22
- p (HC3)
- 0.01
- robust, two-sided
- R²
- 0.039
- SE 0.052 classical vs 0.049 HC3
- Residual Moran's I
- 0.020
- p = 0.26 (permutation)
Spearman's rho of 0.20 (0.04 to 0.36) excludes zero, but the association is weak; the slope's robust interval runs from 0.03 to 0.22 points on the 1-9 scale.
Slope: change in average tone (1-9 scale) per $10,000 of median income. Spearman interval: paired percentile bootstrap (2,000 resamples, seed 57). Slope interval: HC3 heteroskedasticity-robust standard error with a normal reference. A residual Moran's I near zero means the areas' spatial arrangement is not inflating the precision. Each combination of settings is a separate, exploratory comparison; none is corrected for the others.
Spatial statistics: clusters, suppression and the full sensitivity table →
Context
What the 2023 analysis concluded
Revisited with the same data
Pooling suburbs into SA2s makes the claim testable. Across all thresholds the relationship stays weak: Pearson r between 0.00 and 0.22, never explaining more than 5% of the variation. The 2023 reading holds.
| Sentiment of | Min tweets | SA2s | r | ρ | p (r) |
|---|---|---|---|---|---|
| all tweets | ≥ 1 | 279 | 0.11 | 0.10 | p = 0.08 |
| all tweets | ≥ 5 | 229 | 0.21 | 0.19 | p = 0.001 |
| all tweets | ≥ 10 | 201 | 0.22 | 0.22 | p = 0.001 |
| all tweets | ≥ 30 | 147 | 0.20 | 0.20 | p = 0.02 |
| income tweets | ≥ 1 | 167 | 0.14 | 0.10 | p = 0.07 |
| income tweets | ≥ 5 | 85 | 0.17 | 0.18 | p = 0.13 |
| income tweets | ≥ 10 | 69 | 0.08 | 0.09 | p = 0.51 |
| income tweets | ≥ 30 | 33 | 0.00 | 0.00 | p = 0.99 |
Beyond Victoria
Income by capital city area
The SUDO summary grouped all 2,234 SA2s with complete records by Greater Capital City Statistical Area. Its headline figures reproduce exactly: the ACT has the highest median ($60.2k), regional Victoria the lowest median ($41.4k) and lowest mean ($49.6k), and regional Western Australia the highest mean ($71.5k).
Median of SA2 median incomes
- Australian Capital Territory$60.2k
- Rest of NT$57.9k
- Greater Darwin$57.1k
- Greater Perth$52.6k
- Greater Brisbane$51.0k
- Greater Sydney$50.5k
- Greater Melbourne$48.3k
- Rest of WA$47.8k
- Greater Adelaide$46.0k
- Rest of Tas.$45.2k
- Greater Hobart$44.8k
- Rest of Qld$43.9k
- Rest of NSW$43.0k
- Rest of SA$42.9k
- Rest of Vic.$41.4k
Median employee income per job, by industry
- Mining$90.7k
- Electricity gas water waste services$78.5k
- Public administration & safety$56.8k
- Transport postal & warehousing$49.0k
- Manufacturing$48.0k
- Wholesale trade$45.0k
- Construction$40.0k
- Professional scientific technical services$39.8k
- Health care social assistance$39.7k
- Information media telecommunications$37.2k
- Education & training$33.6k
- Finance & insurance services$32.1k
ABS Jobs in Australia 2018-19, the second SUDO dataset on the original dashboard (median across SA2s of each industry's median income per job).
Other platforms
Money talk on Twitter and Mastodon
Twitter's income tweets cluster around neutral. The Mastodon servers looked different in 2023: on mastodon.social most income posts were either clearly positive or clearly negative, which the team read as people showing off or complaining. See the Mastodon page for the full picture.
Twitter, income tweets
93,055 posts
| Score | Count | Share |
|---|---|---|
| 1 (Extremely Negative) | 3055 | 3.3% |
| 2 (Very Strongly Negative) | 4929 | 5.3% |
| 3 (Strongly Negative) | 6316 | 6.8% |
| 4 (Negative) | 6056 | 6.5% |
| 5 (Neutral) | 24640 | 26.5% |
| 6 (Positive) | 9416 | 10.1% |
| 7 (Strongly Positive) | 13362 | 14.4% |
| 8 (Very Strongly Positive) | 13639 | 14.7% |
| 9 (Extremely Positive) | 11642 | 12.5% |
mastodon.social
16,829 posts
| Score | Count | Share |
|---|---|---|
| 1 (Extremely Negative) | 0 | 0.0% |
| 2 (Very Strongly Negative) | 7723 | 45.9% |
| 3 (Strongly Negative) | 0 | 0.0% |
| 4 (Negative) | 0 | 0.0% |
| 5 (Neutral) | 40 | 0.2% |
| 6 (Positive) | 0 | 0.0% |
| 7 (Strongly Positive) | 3861 | 22.9% |
| 8 (Very Strongly Positive) | 1338 | 8.0% |
| 9 (Extremely Positive) | 3867 | 23.0% |
mastodon.au
31,833 posts
| Score | Count | Share |
|---|---|---|
| 1 (Extremely Negative) | 0 | 0.0% |
| 2 (Very Strongly Negative) | 7781 | 24.4% |
| 3 (Strongly Negative) | 2878 | 9.0% |
| 4 (Negative) | 0 | 0.0% |
| 5 (Neutral) | 5075 | 15.9% |
| 6 (Positive) | 2734 | 8.6% |
| 7 (Strongly Positive) | 5680 | 17.8% |
| 8 (Very Strongly Positive) | 4278 | 13.4% |
| 9 (Extremely Positive) | 3407 | 10.7% |
tictoc.social
61,520 posts
| Score | Count | Share |
|---|---|---|
| 1 (Extremely Negative) | 80 | 0.1% |
| 2 (Very Strongly Negative) | 2324 | 3.8% |
| 3 (Strongly Negative) | 3188 | 5.2% |
| 4 (Negative) | 2954 | 4.8% |
| 5 (Neutral) | 17487 | 28.4% |
| 6 (Positive) | 4611 | 7.5% |
| 7 (Strongly Positive) | 7742 | 12.6% |
| 8 (Very Strongly Positive) | 5849 | 9.5% |
| 9 (Extremely Positive) | 17285 | 28.1% |
Method
How this page was built
Tweets were geocoded by the team to 2021 ABS suburbs (SAL) and flagged as income-related by a CouchDB MapReduce view that matched words such as salary, mortgage, afford, job or unfair. Income figures are SUDO personal income for 2015-16 on 2016 SA2 boundaries. To join the two, each suburb was assigned to the SA2 containing its representative point, and tweet counts and score sums were pooled (a weighted average). The IQR filter is the team's own: quartiles from describe().round(2), applied to mean, median and total income in turn. Full details on the data and methods page.