2. Financial Statements

Financial statements are the starting point of company analysis. This tutorial covers how to query DART financial statements, understand the data structure, and extract specific items.

• Querying the 3 core statements (BS, IS, CF)
• Consolidated vs. separate financial statements
• Understanding the year column structure
• Filtering specific accounts
• Quarterly financial statements
• Limitations and caveats of the raw data

Setup

import dartlab

c = dartlab.Company("005930")

The 3 Core Statements

Access the 3 core financial statements as properties on the Company object. These are property accesses, not method calls.

Balance Sheet (BS)

c.BS

The balance sheet shows the balances of assets, liabilities, and equity at a specific point in time. It answers: “As of December 31, what does the company own and owe?”

Core identity: Assets = Liabilities + Equity. If this doesn’t hold, there’s a data issue.

import polars as pl

bs = c.BS
if bs is not None:
    # rows are accounts, columns are years
    print(bs.columns)
    # ['account', '2015', '2016', '2017', ..., '2024']

    # filter for total assets only
    total_assets = bs.filter(pl.col("account") == "자산총계")
    print(total_assets)

Income Statement (IS)

c.IS

The income statement shows revenue and expenses over a period of time. It answers: “How much did the company earn and spend this year?”

is_df = c.IS
if is_df is not None:
    # filter for revenue and operating income
    key_items = is_df.filter(
        pl.col("account").is_in(["매출액", "수익(매출액)", "영업이익"])
    )
    print(key_items)

The same “revenue” may have different account names across companies. Samsung uses “수익(매출액)”, SK Hynix uses “영업수익”. This problem is solved in 3. Time Series.

Cash Flow Statement (CF)

c.CF

The cash flow statement shows how much cash actually came in and went out over a period of time. Even profitable companies can go bankrupt if they lack cash, so this should be read alongside the income statement.

cf = c.CF
if cf is not None:
    cash_items = cf.filter(
        pl.col("account").is_in([
            "영업활동으로인한현금흐름",
            "투자활동으로인한현금흐름",
            "재무활동으로인한현금흐름"
        ])
    )
    print(cash_items)

Consolidated vs. Separate

DART has two types of financial statements:

Samsung Electronics’ consolidated revenue includes subsidiaries like Samsung Display and Samsung SDI. Separate revenue is Samsung Electronics standalone only.

DartLab prioritizes consolidated statements. Only companies without subsidiaries (standalone entities) use separate statements.

# BS, IS, CF automatically prioritize consolidated
c.BS   # consolidated BS (if available) or separate BS

# to check directly in raw data
result = c.fsSummary()
print(f"Consolidated data exists: {result.hasConsolidated}")

Understanding Column Structure

The returned DataFrame columns are account + year columns.

bs = c.BS
print(bs.columns)
# ['account', '2015', '2016', '2017', '2018', '2019', '2020', '2021', '2022', '2023', '2024']

• account: Account name (Korean)
• Year columns: Amount as of that fiscal year-end (in KRW)

Getting Data for a Specific Year

# 2024 total assets
bs = c.BS
row = bs.filter(pl.col("account") == "자산총계")
if row.height > 0:
    value = row["2024"][0]
    print(f"2024 Total Assets: {value:,.0f} KRW")
    print(f"                 = {value/1e12:,.1f} trillion KRW")

Extracting Multiple Items at Once

# extract key BS items
key_accounts = ["자산총계", "유동자산", "비유동자산", "부채총계", "자본총계"]
summary = bs.filter(pl.col("account").is_in(key_accounts))
print(summary)

Quarterly Financial Statements

The default BS/IS/CF are annual data. For quarterly data, use fsSummary().

quarterly = c.fsSummary(period="q")

quarterly.BS   # quarterly balance sheet
quarterly.IS   # quarterly income statement
quarterly.CF   # quarterly cash flow statement

Quarterly column names use the format 2020Q1, 2020Q2, 2020Q3, 2020Q4.

qbs = quarterly.BS
print(qbs.columns)
# ['account', '2020Q1', '2020Q2', '2020Q3', '2020Q4', '2021Q1', ...]

Quarterly income statement (IS) values are cumulative. The Q2 column for revenue is not “Q2 revenue” but “Q1+Q2 cumulative revenue”. For standalone quarterly revenue, use the time series engine in 3. Time Series.

Caveats

Account names differ across companies

The same “revenue” may have different names at different companies. Filtering with pl.col("account") == "매출액" may return empty results for some companies.

# this won't work for all companies
bs.filter(pl.col("account") == "매출액")   # may not find in some companies

# use the time series engine to solve this
# → see 3. Time Series

Amount unit

DART raw data amounts are in KRW (Korean Won). Values returned by BS/IS/CF properties are also in KRW.

# Samsung total assets is 400 trillion, not 4 trillion
# → unit is KRW, so 400,000,000,000,000 (400T KRW)

None return

Companies without data return None. Always check for None.

bs = c.BS
if bs is not None:
    print(bs)
else:
    print("No balance sheet data available")

Next Steps

• 3. Time Series — Account standardization, standalone quarterly time series, TTM
• 4. Ratios — ROE, debt ratio, FCF, and more auto-calculated