# Custom Numeric Chart A custom-numeric chart puts any numeric quantity on the horizontal axis. Unlike the candlestick / line / area family (which require time on the x-axis) and unlike the yield-curve chart (which expects months) and the options chart (which expects strikes), `tvl.custom_numeric` makes no assumption about what your x-values *mean*. Internally `tvl.custom_numeric` uses the same `createOptionsChart` renderer as `tvl.options_chart`. Use it whenever your x-axis is a generic numeric quantity (frequency in Hz, distance in km, price level for a profile chart, basket index) rather than time, maturity, or strike. ## What are custom-numeric charts useful for? - **Frequency / Bode plots**: Magnitude or phase vs. frequency on a numeric axis with no time semantics. - **Distance / depth profiles**: Price vs. depth-of-book level, density vs. depth, anything indexed by a numeric position. - **Price-level profiles**: Volume profile, market profile (TPO), and any other “histogram by price level” rendering. - **Custom indices**: Basket weights vs. constituent number, regression coefficients vs. feature index: any “value vs. integer position” view. ## Examples ### A basic line on a numeric x-axis The chart treats `x` as a numeric coordinate, so any numeric column will do. Pass it explicitly along with the value column. Build a small synthetic table inline by attaching an integer `X` to the `tvl.data.values()` series. ```python order=custom_numeric,profile import deephaven.plot.tradingview_lightweight as tvl profile = tvl.data.values().update(["X = (double)Index"]) custom_numeric = tvl.custom_numeric(profile, x="X", value="Value") ``` The line traces `Value` over the integer x-axis. Reuse a strike, tenor, or any other numeric column verbatim when you want a generic numeric rendering of data you might otherwise feed to `options_chart` or `yield_curve`. > [!NOTE] > Viewport-aware downsampling currently only applies to time-axis charts (`chart_type == "standard"`). Custom-numeric charts ship raw rows to the client, so keep the input table modest (tens of thousands of rows) or pre-aggregate server-side. ### Switch to an area rendering `series_type="area"` swaps the line for a filled area. Combine it with `top_color`, `bottom_color`, and `line_color` for a custom gradient. ```python order=custom_numeric,profile import deephaven.plot.tradingview_lightweight as tvl profile = tvl.data.values().update(["X = (double)Index"]) custom_numeric = tvl.custom_numeric( profile, x="X", value="Value", series_type="area", top_color="rgba(38, 166, 154, 0.45)", bottom_color="rgba(38, 166, 154, 0.0)", line_color="#26a69a", ) ``` The gradient fades to transparent at the bottom. ### Switch to a histogram rendering `series_type="histogram"` renders one bar per row, a natural fit for volume-profile and market-profile views where the x-axis is a price level. ```python order=custom_numeric,profile import deephaven.plot.tradingview_lightweight as tvl profile = tvl.data.values().update(["X = (double)Index"]) custom_numeric = tvl.custom_numeric( profile, x="X", value="Value", series_type="histogram", color="#9467bd", ) ``` A purple bar rises at each integer x-value. ### Customize color, width, and title In line and area modes, `color`, `line_width`, and `title` work the same way as in the other chart types. `line_width` is an integer pixel width 1-4. ```python order=custom_numeric,profile import deephaven.plot.tradingview_lightweight as tvl profile = tvl.data.values().update(["X = (double)Index"]) custom_numeric = tvl.custom_numeric( profile, x="X", value="Value", color="#1f77b4", line_width=3, title="Custom profile", ) ``` The line is blue, three pixels wide, with a custom legend entry. ## API Reference Use this when the x-axis represents arbitrary numeric values (frequency, distance, price level, etc.) rather than timestamps. Internally uses the same createOptionsChart renderer as options_chart(), which provides a linearly-spaced numeric horizontal scale. **Returns:** `TvlChart` A chart with a generic numeric x-axis. **Raises:** ValueError -- If series_type is not "line", "area", or "histogram".