How to Use Market Data in a High School Finance Research Project
Market data can turn a finance paper from a collection of opinions into a testable argument. It can also create false confidence when the units, dates, or definitions are misunderstood.
The goal is not to collect the largest possible spreadsheet. The goal is to choose evidence that directly helps answer a focused research question.
Begin with the claim you want to test
Before downloading anything, write the possible claim in one sentence.
For example: “Unexpected inflation increased short-term government-bond yields more than long-term yields.”
That claim tells you that you need an inflation measure, a way to identify surprises, yields at two maturities, and a defined period. If a dataset does not help test the claim or a competing explanation, it may not belong in the project.
Understand the main types of finance data
Market-price data
This includes share prices, bond prices and yields, exchange rates, commodity prices, credit spreads, and market indexes. Prices reflect expectations as well as current conditions, so they may move before an event occurs.
Use consistent trading dates and state whether values are daily closes, weekly averages, or monthly observations.
Macroeconomic data
Inflation, GDP growth, unemployment, trade balances, fiscal balances, policy rates, and foreign-exchange reserves provide economic context.
Macroeconomic series are often revised. Record the release or download date and consider whether your question requires the value known at the time or the latest revised estimate.
Company data
Revenue, profit margins, debt, cash flow, capital expenditure, and segment results come from financial statements and regulatory filings.
Check whether the company uses a calendar or non-calendar fiscal year. Do not compare a quarterly figure with a full-year number or mix currencies without conversion.
Policy and event data
Central-bank decisions, budget announcements, elections, rating changes, defaults, and regulatory actions can define the events around which you analyse market behaviour.
Record the exact announcement time when using daily or intraday market data. A decision announced after markets close may appear in the following trading day’s price.
Use reliable sources
Strong projects begin with sources that define their variables clearly. Depending on the question, useful starting points include:
Central banks and national statistical agencies
Government finance ministries and debt-management offices
Company annual reports and regulatory filings
The World Bank, International Monetary Fund, and OECD
FRED and other official economic-data portals
Exchange or index-provider documentation
Credible market-data services available through your school
Avoid copying a number from an image, social-media post, or search-result snippet when the original dataset is available.
Keep a data dictionary with the series name, source, unit, frequency, date range, and transformation. Future you should be able to understand every column.
Check units before calculating
Many research errors are unit errors.
A move from 4% to 5% is an increase of one percentage point, not one percent. A bond yield quoted as 6.25 may mean 6.25%, while an exchange rate might be expressed as local currency per US dollar or the reverse.
Decide whether a series is nominal or adjusted for inflation. Confirm whether debt figures are in local currency, US dollars, or a percentage of GDP. Check whether large numbers are reported in thousands, millions, or billions.
Write the unit into every column heading.
Align dates and frequencies carefully
Two series cannot be compared responsibly until their dates align.
If one series is daily and another is monthly, decide whether to convert daily data into an end-of-month value or an average. The choice should match the question. An end-of-month exchange rate may suit a balance-sheet comparison, while an average may suit a monthly trade analysis.
Do not fill every missing observation automatically. Markets close on holidays, some indicators are released quarterly, and missing values can carry information about how the dataset was constructed.
Build a clean research table
Keep raw data unchanged on one sheet and perform cleaning on another. A working table might include:
Date
Main outcome
Explanatory variable
Comparison or benchmark
Event flag
Source note
Calculation or transformation
Use simple, auditable formulas. If you calculate a percentage change, record the formula once and apply it consistently.
Name files and tabs clearly. “Final_final_v3” is not a research system.
Create charts that reveal the argument
A chart should answer a specific question.
For a central-bank study, plot the policy rate and the relevant bond yield with decision dates marked. For a currency-crisis study, compare reserves, the exchange rate, and the sovereign spread using separate panels or carefully explained scales.
Every chart needs:
A descriptive title
Clearly labelled axes
Units
A defined period
A source note
An explanation in the text
Avoid three-dimensional charts, decorative effects, and truncated axes that exaggerate small changes.
Distinguish correlation from mechanism
Two variables moving together does not prove that one caused the other.
Ask what mechanism would connect them. Consider whether a third factor could move both. Global interest rates, commodity prices, political news, and general risk sentiment frequently affect several financial variables at once.
A strong report states: “The pattern is consistent with this explanation, but the comparison cannot fully rule out these alternatives.”
That sentence is more credible than claiming certainty the method does not provide.
Document every transformation
If you convert currencies, calculate returns, create an index, or remove outliers, explain exactly what you did.
A reader should be able to reproduce the result using the original source. Keep formulas visible, record assumptions, and avoid manually overwriting calculated cells.
When a result changes substantially after a reasonable alternative calculation, report that sensitivity rather than hiding it.
Know when the data is not enough
Sometimes the correct conclusion is that the evidence is inconclusive.
A short sample, inconsistent definitions, missing dates, or one unusual event may prevent a strong claim. Identifying that limitation is part of the research, not a failure.
Explain what additional data or comparison would make the answer stronger.
Turn the dataset into a report
Data becomes research only when it supports a clear argument. Move from the table to an evidence outline: claim, chart or calculation, interpretation, competing explanation, and limitation.
The GRF Research Desk helps five high-school Fellows move from a focused finance question to a complete, individually authored report over six weeks. Learn about the September cohort at https://www.globalresearchfellowship.com/research-cohort?src=blog-market-data-guide-end