What Financial Data Can Show
A figure is a measurement with a definition. An indicator describes. A model can summarise and can also drop the definition. This note is a reading habit, not a method for choosing an asset.
Financial pages are full of numbers that look finished. A percent, an index level, a chart with no sentence under it. The finish is visual. Under it, every figure was built: someone chose what to count, over what period, in what unit, and what to leave out. Data literacy is the habit of asking for that construction before treating the figure as a fact about your own list.
The habit is for beginners, including readers in Canada who meet national statistics, market quotes, and automated summaries in the same week. It does not rank indicators, and it does not say which figure should drive a decision. Decisions are outside the page. The page only slows the number down.
A figure is a measurement with a definition
A useful figure can be restated as a sentence with a verb of counting. “The consumer price index rose” is incomplete until you can say whose basket, which geography, which month compared with which earlier month, and whether the agency has warned that the month will be revised. Statistics Canada publishes those definitions because the number is not self-explanatory. Reading the definition is not the same as agreeing with a political story about the number. It is how you learn what was counted.
Market figures have the same need. A closing price belongs to a venue and a session, as the markets note described. A fund’s reported value belongs to a method and a time. A yield belongs to a formula. If you cannot point to the sentence that built the figure, you are looking at a label, not a measurement. Labels travel well on screens. Measurements travel with footnotes.
Indicators describe, they do not instruct
An indicator compresses a wider set of figures into one series: inflation, unemployment, a market index, a ratio on a company report. Compression is helpful and lossy. The series can describe a history. It cannot tell you what to do about the next period, because the next period is not in the series. People slide from “this moved” to “so you should” in a single paragraph. The slide is the thing to mark.
Company and market indicators are especially easy to over-read because they sit next to prices. A ratio is not a forecast and not a verdict that a price is wrong. It is a comparison inside a definition. Two firms can share a ratio and differ in the debts, the currency, and the customers that the ratio averaged away. The balance-sheet station matters here: if you do not know which stock or flow the indicator describes, you cannot place it on a household list either.
Past movement in an indicator is not a future result. That sentence is dull and it is the one indicators most often violate in popular retelling. This site does not draw those movements as charts, because a line that rises is too easily stolen as a promise. The absence of a chart is deliberate.
Alternative data and the missing context
Besides official series, commentary now cites card spending, satellite counts, app downloads, and other traces. These can be measurements too, if the counting rule is stated. Often the rule is proprietary, the sample is unnamed, and the link to a household or a firm is assumed. Alternative does not mean sharper. It means the definition is somewhere else, sometimes somewhere you cannot read.
The educational response is not to sneer and not to adopt. Ask the same three questions you would ask of a statistics-agency release, and notice when the answers are missing. A missing base period, a missing geography, or a missing statement of who is in the sample is a limit on what the figure can show. The limit is not a reason to invent a story that fills the gap.
When a model writes the summary
Models, including large language models, can summarise a release, a filing, or a thread of commentary. The summary can be fluent and still drop the definition, the revision note, or the unit. Fluency is not a source. A model does not become a witness because it writes in complete sentences. It is a compression tool with a different lossiness from an indicator: it can omit the caveat that was the most important line.
Use a summary, if you use one at all, as a table of contents. Then open the document it claims to describe. If the document is a regulator’s explainer or a statistical release, the primary text is public. If the summary cannot point to that text, it is not yet a measurement. This station discusses models as a reading problem. It does not offer a model that suggests trades, and it does not claim that automation makes financial judgement easier. It often makes the missing definition harder to notice.
A three-part reading habit
The habit is short enough to reuse. It applies to a national statistic, a market quote, an indicator, and a paragraph written by software.
- 01 Definition. What was counted, in which unit, for whom?
- 02 Base. Compared with what earlier period, basket, or rule — and can that base be revised?
- 03 Limit. What can this figure not say, including anything about your own list or about a future price?
Bring the habit back to the first station when you are tempted to paste a national average onto a personal balance sheet. Averages are about groups. Your list is a photograph of one household, and this project does not have that photograph. Bring it to the markets note when a price has no venue and no currency written beside it. The editorial team of Vertex Vest offers the habit as education. Markets can still move against a careful reader. Care is not protection. It is the difference between a number you can question and a number that only looks finished.