CSV format
framechart accepts standard UTF-8 CSV files. The first row must be a header. Values can be quoted or unquoted; commas are the only supported delimiter.
Long format vs wide format
framechart auto-detects which shape your CSV is in. You can override the detection with the Data Mapping group in the Inspector.
Long format (recommended for races)
One row per (time × entity) pair. Easy to add new time steps.
Year,Country,GDP 2010,United States,15049 2010,China,6101 2010,Germany,3310 2015,United States,18206 2015,China,11016 2015,Germany,3355
Wide format
One row per entity; time steps are column headers. Common in Excel exports.
Country,2010,2015,2020,2023 United States,15049,18206,21354,27357 China,6101,11016,15103,17795 Germany,3310,3355,3937,4456
Both shapes work for Bar Race and Line Race. framechart reads the shape and normalises to long format internally. If detection goes wrong, use the Data Mapping pickers to explicitly assign each column role.
Line chart format
Long format: an X column (time or category), a Y column (numeric), and an optional series column (entity names for multi-line charts).
Date,Company,Revenue 2020-Q1,Apple,58313 2020-Q1,Microsoft,35021 2020-Q2,Apple,59685 2020-Q2,Microsoft,38033
Wide format also works: each numeric column becomes one line. The first column is treated as the X axis; every other column is a separate series.
Table format
Any tabular CSV. Column headers become the table header row. Every column is displayed by default; use the Columns group in the Inspector to hide individual ones.
Rank,Country,GDP (T USD),Growth 1,United States,27.4,2.5% 2,China,17.8,5.2% 3,Germany,4.5,0.1%
Numeric detection
A column is treated as numeric if every non-header cell parses as a floating-point number
(after stripping leading/trailing whitespace, optional commas as thousands separators, and
an optional leading $ or trailing %).
27357→ numeric27,357→ numeric (thousands separator stripped)$27,357→ numeric2.5%→ numeric (% stripped, value 2.5)United States→ text
Date / time columns
When the X axis mode is set to Auto or Time, framechart tries to parse the X column values as dates. Supported formats:
2024— year integer2024-03— year-month2024-03-15— ISO 8601 date2024-Q1— year + quarter
Parsed dates control proportional spacing on the X axis. If parsing fails, framechart falls back to categorical (evenly-spaced) layout.
Data mapping overrides
Auto-detection is a heuristic: it picks the first text column as the category and the last numeric column as the value, for example. When your CSV has a different structure, use the Data Mapping group in the Inspector to explicitly assign each column to its role (Category, Value, X, Y, Series, Time). Option 0 in every picker is Auto — switch back to it to re-enable heuristics.