Skip to content
Back to the metric guideGuide

GUIDE · Learning path step 28 / 40 · Reading

GUIDE · Step 28 / 40

What free data Jupick is built on: a map of data sources and their limits

Jupick uses no paid data: Korean prices come from Korea Exchange (KRX), Korean financials from DART, overseas stocks from Yahoo Finance, crypto from CoinGecko, macro indicators from FRED and news from RSS headlines. Each source has its limits.

Jupick uses no paid data

Jupick takes stocks, indices, crypto, commodities, exchange rates and macro indicators only from data that is freely published. It does not use paid data subscriptions or sources that require an issued key. This principle runs in the same vein as the service itself being free.

Because it uses only free sources, Jupick cannot get everything from one place. Each kind of asset is supplied well by a different source, so the sources split into several. Several sources also mean several sets of circumstances and limits.

This article lays out, like a single map, which source Jupick gets what from and how far each source can speak. Knowing where a number on screen came from also helps gauge how far to trust it.

Korean prices and market caps come from Korea Exchange

Share prices, price changes and market caps for Korean stocks come from the full list of listed stocks published by Korea Exchange (KRX), fetched in one go. Rather than looking up each stock separately, Jupick takes the whole list at once and splits it up, so it can fill in prices for the major Korean stocks quickly.

At one time another free tool also supplied Korean financials, but that tool changed to require a member login and no longer returns values. Free sources can close their doors like this one day. So prices now come from the exchange, and financials from a different place, described next.

Korean financials come from DART, overseas stocks from Yahoo

Financial metrics for Korean stocks come from DART, the electronic disclosure system of Korea's Financial Supervisory Service. Jupick reads the financial statements companies have disclosed and calculates metrics such as ROE and debt ratio. Because disclosures are the source, a value the company has not yet disclosed may be blank.

Overseas stocks come from Yahoo Finance. Because Korean and overseas financials come from different sources, even metrics with the same name may differ in the definitions or reference dates used in the calculation. That is why comparing the same company with another site sometimes yields different values.

Either way, there are days when a data source goes silent. Jupick rewrites each document entirely on every update, so leaving things as they are on a day when the source returns no value would erase facts that were there. So on such days it restores the value from the previous update to keep blanks from spreading.

Indices, exchange rates and commodities come in twice; correlations use closing prices

Indices such as KOSPI and KOSDAQ, along with exchange rates and commodities, come from the FinanceDataReader library. Because one source once stopped quietly for the Korean indices, Jupick also fetches the same indices from Yahoo and uses whichever value is more recent.

The asset class correlation table takes the last 90 days of closing prices for these indices, exchange rates and commodities and calculates how closely each pair moved together. It is built from one source's closing prices only, so if that source stops, the table stops too. That is why it always notes that correlation is not a prediction but a record of the past 90 days.

Crypto from CoinGecko, macro indicators from FRED

Crypto prices and grades come from CoinGecko's free data. Crypto has no financial statements, so its grades use a yardstick entirely different from stocks. A stock S and a crypto S do not mean the same thing.

Macro indicators such as interest rates, inflation and unemployment come from FRED, the public time series of the Federal Reserve Bank of St. Louis. It is a public CSV, so no issued key is needed. Macro indicators carry neither grades nor predictions; Jupick shows only the facts of the current value and its change from the past.

News: headlines only

Market news comes from the RSS feeds of several outlets, but Jupick stores neither article bodies nor summaries, only headlines and links. This reduces the copyright burden. For the full text, the link sends readers to the original outlet.

The positive, negative and neutral badges on each headline are assigned from the words of the headline alone. They are not judgments made after reading the article; they are the result of sorting each headline's tone with a predefined Korean keyword dictionary. So the label reflects the coloring of the headline's vocabulary, not the direction of the share price.

The limits of free data, and what Jupick does about them

Free sources are not perfect. One day they close their doors, stop quietly, or leave values out. When that happens Jupick does not make up missing values. Instead it uses redundancy to choose the side that has not stopped, restores the previous value to keep blanks from spreading, and for values it still cannot fill, states the observation date and leaves them honestly as they are.

Having several sources means having several limits, and not hiding those limits is how free data keeps trust. That is why notes throughout this site state, for each number on screen, where it came from and how far it can speak.

This content is for information only and is not investment advice. Jupick does not forecast future prices and gives no trading advice of any kind. Investment decisions and responsibility for them rest with the user.

This content is for information only and is not investment advice. Investment decisions and responsibility for them rest with the user.