A topic chart is only as useful as the readings behind it. A missing value, a reversed date sequence or a second copy of the same record can change what you see before you have even started deciding which subject to write about.
Shane’s Topics lets you import your own research as CSV, compare the supplied readings and create a research outline. Preparing a small, checked file first makes that process easier to trust. The tool does not collect live search data or verify the source for you.
1. Give each series a clear identity
Start with one topic, one source, one region and one reporting window per row. Keep the original source file separately so you can check the numbers later. In your notes, identify the measurement: a relative-interest index, an estimated search count or another clearly defined unit.
Use a topic name that distinguishes each series. For example, “Refillable cleaners | US | Jan–Mar 2026” and “Refillable cleaners | UK | Jan–Mar 2026” are easier to tell apart than two records both called “Refillable cleaners.” You still have separate Region and Source fields; the longer name helps when selecting a record for a brief.
This matters in the current outline builder. It looks up the source readings using the first topic record whose name matches, ignoring letter case. Two records with the same name can therefore produce a brief whose source readings come from the earlier record, even when you selected the later card. Distinct names and a final check of the outline avoid that mix-up.
2. Check every period-and-reading pair
Lay out the source data in two columns before compressing it into a CSV cell: period on the left, reading on the right. Sort the pairs together, oldest to newest. The tool keeps the order you supply; it does not interpret and reorder the date labels.
Use at least two readings, with exactly one period label for each. A row containing January, February and March should have three corresponding values. Keep the same reporting interval within a series and explain any gaps. The chart spaces observations evenly, so a large gap in calendar time does not create a larger horizontal gap between points.
Only include readings you can support. The current importer converts an empty reading between separators to zero. It also drops text that cannot be read as a number and negative readings. These behaviors can turn incomplete input into a different series.
- A blank is not a measured zero. The values
12||27become 12, 0 and 27. Use zero only when the source actually reports zero. - Text is not a missing-data marker. In
12|unknown|27, the word is dropped. With three period labels, the resulting count mismatch is rejected. - Matching counts do not prove correct alignment. If both a label and a reading are missing or removed in different positions, a row can still have equal counts while pairing the wrong values with the remaining periods.
Resolve missing observations in your source sheet. If you deliberately omit an unavailable period, omit its matching reading too and explain the gap in Notes. Do not invent a value to complete a smooth-looking line.
3. Build a small CSV with the expected columns
The required columns are topic, periods and values. You can also include category, region, source and notes. The CSV template button provides a starting file. Replace the example row before importing your research.
The following fictional example shows one complete record. Its values are invented to demonstrate the format, not evidence of a real trend.
topic,category,region,source,periods,values,notes
"Refillable cleaners - US - Jan-Mar 2026",Other,"United States",Other,"2026-01|2026-02|2026-03","12|18|27","Fictional example only; replace with verified source readings."Vertical bars separate the periods and readings within their cells. The tool also accepts commas or semicolons inside those cells. If you use commas within a CSV cell, enclose the entire cell in double quotes so the commas do not become extra columns.
Let a spreadsheet’s CSV export handle quoting where possible. If you edit the file as plain text, double any quotation mark inside a quoted cell. A note containing the words source says "index" should appear as "source says ""index""" in the CSV. Open the final file once to check that each topic remains one row with the expected fields.
4. Import once, then compare the result with the file
Before an import, use Export CSV to save the records already in your workspace. Keep the new input file separate from that export. Importing appends records; it does not update a matching topic, merge repeated series or replace the workspace. Importing the same file again creates another copy.
Start with a small batch that is easy to inspect. After a successful import, confirm the reported number of added records, clear any search or category filter, and check the new cards. Compare the latest reading, number of readings and total change with your source file.
For the fictional example above, the last reading is 27 and the card has three readings. The change from 12 to 27 is 125%, which receives the Rising label. That arithmetic checks the example’s endpoints; it does not establish demand or explain why a real series changed.
Use Export CSV again to inspect the complete stored period and value lists. Check those pairs against the input file, especially when the endpoints look right but an intermediate observation may be wrong. Export covers all topic records currently loaded in the page, even when a filter hides some cards.
5. Stop and reset the session after an import error
A failed batch needs extra care. The current importer processes rows in order. If an early row is valid and a later row fails, the valid earlier row can remain in the open page’s memory even though the import did not finish and the dashboard has not refreshed. Retrying the corrected batch in that same session can duplicate the earlier row and save both copies.
If an import fails, note the error and pause before clicking other controls. Copy any unfinished outline or form text into your own file, then reload the page before retrying. Reloading returns to the last successfully saved workspace, provided no later action has saved the partial import. Fix and review the whole CSV outside the tool, then import it once.
If you have already retried or used another control after the error, do not assume the earlier workspace is intact. Export the current records and compare them with your pre-import copy before using the totals. Re-importing that backup into the same workspace would append its rows again, so it is not a replacement or undo operation.
6. Verify the brief before writing the article
Once the records are checked, choose Create brief on the intended card. Read the topic name, region, source, date range and Source readings line together. They should all describe the same series. Keep the source file and the finished outline with the project so another person can follow the reasoning.
The useful outcome is a research question you can investigate: what changed, what evidence supports it, and whether the topic fits your audience. A clean import gives that decision a dependable starting point without turning observed interest into a prediction.