Export data
Download call-level measures and keep the query and settings behind them.
Download the main file
Open a completed dataset, or find it in My Library. Open Downloads and, under Firm panel data, choose Download CSV beside Metrics (for earnings calls) or Filing metrics (for SEC filings).
The downloaded filename ends in _firmlevel.csv. Its prefix includes a timestamp, the dataset name, and an identifier.
Each row is one earnings call in the selected date range, identified by earningscallID. The file contains your topic measures and supporting fields such as company, date, and transcript length.
Keep zero rows. Calls with no query matches are still part of the sample. Removing them changes averages and aggregate statistics. See what zero rows mean.
Choose any additional files
In Downloads, choose Download CSV beside the file you need. Available downloads depend on the dataset, context, and plan.
- Overall risk and sentiment: counts across each call, without requiring a topic match.
- Crosswalk to GVKey: a crosswalk for Compustat joins. Verify the matches before using them.
- Keyword breakdown: per-term counts when the breakdown option is included.
- Time-series data: quarterly aggregate series under Time-series data (Quarterly time series (all measures) or Quarterly exposure time series).
- Weighted query terms: the selected terms and weights for a weighted dictionary query. Find this under Search details.
Save the query and settings
Under Search details in Downloads, choose Download JSON beside Search configuration, where available. The filename includes the dataset name and identifier and ends in -config.json.
The JSON records the engine, matching context, and scoring choices as well as the query and filters. Keep it with your CSV files and complete the research record.
Before using the export for research, check the date range, matched sentences, and sample availability and topic variation.
Next, join and normalize the output for your analysis.
Column reference
The headings below match the labels in Downloads. Saved filenames have dataset-specific prefixes; use the filename endings to identify them.
Metrics / Filing metrics (_firmlevel.csv)
Filename ending: _firmlevel.csv (labeled Metrics for earnings calls and Filing metrics for SEC filings).
Columns vary with the dataset generation and settings. Use start_time for the call date and time; date and date_q may be absent.
| Column | Definition |
|---|---|
earningscallID | Unique identifier for the earnings call. Use it for joins, deduplication, and citations. |
company_name | English company name. |
date | Date of the earnings call, based on UTC time, when included. |
event_name | Name of the earnings-call event, usually including the fiscal period discussed. |
start_time | ISO 8601 timestamp for when the call started, in UTC. Use it when precise timing matters. |
exposure, risk, positive, negative, sentiment | The topic-conditioned measures. See measure definitions. |
nr_of_sentences | Total number of sentences in the full transcript. |
nr_of_words | Word count supplied with the transcript metadata. |
tickersymbol | Ticker symbol. |
permid | Permanent entity identifier, where available in the corpus metadata. |
ric | Instrument code, where available in the corpus metadata. |
isin | International Securities Identification Number. |
cik | SEC Central Index Key, mainly for U.S. entities. |
date_q | First day of the call's calendar quarter, when included. Calendar quarter, not fiscal quarter. |
headquarterscountry | Registered headquarters country. |
hqcountrycode | Two-letter headquarters-country code, where available. |
trbc | Business classification code. |
economic_sector | Label for the first two digits of trbc. |
business_sector | Label for the third and fourth digits of trbc. |
nr_of_sentences_filtered | Number of sentences passing the section, speaker-affiliation, or speaker-title restrictions. Present only when such a restriction was used. |
matched_speaker_names | Names of speakers whose sentences matched, populated when one of the six speaker-title presets (CEO, CFO, COO, CTO, Analyst, Operator) was used. |
The file can also contain columns for measures under development, which these docs do not define. Save the query and settings with the export to preserve its scoring context.
Identifier columns other than earningscallID can be empty for some calls. Check identifier completeness in your target sample before designing a join around one of them.
Overall risk and sentiment
Filename ending: _unconditional.csv.
Overall risk and sentiment counts, not restricted to the query. One row per earnings call.
| Column | Definition |
|---|---|
earningscallID | Unique identifier for the earnings call. |
nr_of_sentences | Total number of sentences in the full transcript. |
unconditional_positive_sentences_count | Sentences with at least one positive sentiment word. |
unconditional_negative_sentences_count | Sentences with at least one negative sentiment word. |
unconditional_risk_sentences_count | Sentences with at least one risk or uncertainty synonym. |
unconditional_sentiment | Overall Positive Sentiment minus Overall Negative Sentiment, included when sentiment is among the selected measures. |
Join it to the Metrics file on earningscallID to express topic-conditioned measures relative to all risk or sentiment language in the call. See Overall risk and sentiment for the definitions.
Crosswalk to GVKey
Filename ending: _gvkeys.csv.
A convenience crosswalk for Compustat joins. One row per earnings call.
| Column | Definition |
|---|---|
earningscallID | Unique identifier for the earnings call. |
gvkey_compustat | Compustat GVKey, where a match exists. |
sic_compustat | Four-digit SIC code, where available. |
sic_compustat_description | Label for the SIC code. |
Treat the crosswalk as an input to verify, not as a guaranteed match: check match quality in the downstream dataset and document the checks you performed.
Keyword breakdown
Filename ending: _keyword_breakdown.csv.
Produced when Download counts for each search term is selected under Options on the previous search engine. See engine-dependent features. One row per keyword and earnings call.
| Column | Definition |
|---|---|
term | The query keyword or phrase. |
earningscallID | Unique identifier for the earnings call. |
exposure_count, risk_count, positive_count, negative_count | Sentence counts for this term's matches in this call. |
There is no per-term sentiment column; calculate it as positive_count - negative_count if needed. Additional columns for measures under development are outside this reference.
Use it to inspect individual terms, such as whether a broad keyword accounts for many matches. A sentence can match several terms, so their counts can overlap. Adding or subtracting per-keyword counts does not reliably reconstruct the combined measure. To measure a revised query, edit and rerun it, then review the new dataset.
Time-series data
Filename ending: _timeseries.csv or _timeseriesq.csv.
Under Time-series data in Downloads, choose Download CSV beside Quarterly time series (all measures) (or Quarterly exposure time series for Exposure-only datasets such as SEC filings). This file exports the aggregate quarterly series shown in the results chart.
Weighted query terms
Filename ending: _weighted_terms.csv.
This file lists the selected dictionary terms and their raw and normalized weights, one row per term. Keep it with the other downloads.
Interpret the output using the saved scoring settings and the applicable weighted-measure methodology.