Sample CSV kits
The same markets as the sample shelf. Open a market for its files and steps.
A ratings study uses profiles.csv and attributes.csv.
Column guide.
Transportation 1 sample
Train commute
People score each offer
2 attributes (ticket price, platform wait), 9 profiles, choice set for share simulation.
Steps
- Download the zip (or at least profiles.csv and attributes.csv).
- Open profiles.csv — change only the rating column (0–100) for each row.
- Upload via Studio → Study files, or run the bundled template from Examples.
- Open Report for market share and attribute importance charts.
| File | Download |
|---|---|
attributes.csv |
attributes.csv |
choice_set.csv |
choice_set.csv |
choice_tasks.csv |
choice_tasks.csv |
profiles.csv |
profiles.csv |
settings.csv |
settings.csv |
Food & beverage 1 sample
Coffee subscription
People score each offer
Monthly price vs delivery speed — 9 profiles with share simulation.
Steps
- Download the Coffee subscription zip.
- Edit ratings in profiles.csv to reflect your taste (0–100 per bundle).
- Run bundled from Examples, or upload your edited CSVs under Study files.
- Compare plans in the market share chart on Report.
| File | Download |
|---|---|
attributes.csv |
attributes.csv |
choice_set.csv |
choice_set.csv |
choice_tasks.csv |
choice_tasks.csv |
profiles.csv |
profiles.csv |
settings.csv |
settings.csv |
Travel & hospitality 1 sample
Hotel weekend
People score each offer
Nightly rate vs distance to city center — weekend stay trade-offs.
Steps
- Download the Hotel weekend kit.
- Adjust rating scores for each room/distance combo in profiles.csv.
- Upload under Study files or run the bundled template from Examples.
- Use Report to see which weekend option wins share.
| File | Download |
|---|---|
attributes.csv |
attributes.csv |
choice_set.csv |
choice_set.csv |
choice_tasks.csv |
choice_tasks.csv |
profiles.csv |
profiles.csv |
settings.csv |
settings.csv |
Media & entertainment 1 sample
Streaming plan
People score each offer
Monthly fee vs ad load — fictional streaming tiers with share simulation.
Steps
- Download the Streaming plan zip.
- Set ratings in profiles.csv for each fee/ad combination.
- Run from Examples or upload via Study files.
- Review share simulation on Report.
| File | Download |
|---|---|
attributes.csv |
attributes.csv |
choice_set.csv |
choice_set.csv |
choice_tasks.csv |
choice_tasks.csv |
profiles.csv |
profiles.csv |
settings.csv |
settings.csv |
Consumer electronics 3 samples
Blueberry phones
People score each offer
5 attributes, 32 profiles, catalog + shortlist for purchase ranking.
Steps
- Download profiles.csv, attributes.csv, and catalog.csv.
- Edit ratings in profiles.csv; optional: adjust catalog.csv SKUs.
- Upload via Study files or run the bundled template.
- Open Report for part-worths and purchase ranking.
| File | Download |
|---|---|
attributes.csv |
attributes.csv |
catalog.csv |
catalog.csv |
profiles.csv |
profiles.csv |
settings.csv |
settings.csv |
shortlist.csv |
shortlist.csv |
Laptop purchase
People score each offer
5 attributes (brand, form factor, RAM, screen, storage) — 32-profile ratings study.
Steps
- Download profiles.csv, attributes.csv, and catalog.csv.
- Edit ratings per laptop configuration in profiles.csv.
- Run bundled from Examples or upload via Study files.
- Open Report for utilities and purchase ranking.
| File | Download |
|---|---|
attributes.csv |
attributes.csv |
catalog.csv |
catalog.csv |
profiles.csv |
profiles.csv |
settings.csv |
settings.csv |
shortlist.csv |
shortlist.csv |
Laptop purchase (choice)
People pick one offer
Choice-based conjoint on laptop configurations — 36 tasks, demo picks included.
Steps
- Download the Laptop purchase CBC kit.
- Review choice_observations.csv.
- Run bundled from the sample shelf.
- Inspect coefficients and fit on Report.
| File | Download |
|---|---|
attributes.csv |
attributes.csv |
choice_observations.csv |
choice_observations.csv |
choice_tasks.csv |
choice_tasks.csv |
settings.csv |
settings.csv |
Vehicles 6 samples
Metro auto market
People score each offer
10 respondents, panel max-utility market share across scenarios.
Steps
- Download the panel kit zip.
- Panel studies: keep respondent structure unless you know the format.
- Run bundled from Examples — pick a market scenario on Report.
- Compare current vs new-product share on Report.
| File | Download |
|---|---|
attributes.csv |
attributes.csv |
choice_set.csv |
choice_set.csv |
choice_set_scenario1.csv |
choice_set_scenario1.csv |
choice_set_scenario2.csv |
choice_set_scenario2.csv |
partworth_map.csv |
partworth_map.csv |
respondents.csv |
respondents.csv |
settings.csv |
settings.csv |
Bike purchase
People score each offer
E-bike price vs battery range — ratings conjoint with share simulation.
Steps
- Download the Bike purchase zip.
- Edit ratings in profiles.csv for each price/range combo.
- Run bundled from Examples (Ratings) or upload via Study files.
- See which bike configuration wins share on Report.
| File | Download |
|---|---|
attributes.csv |
attributes.csv |
choice_set.csv |
choice_set.csv |
choice_tasks.csv |
choice_tasks.csv |
profiles.csv |
profiles.csv |
settings.csv |
settings.csv |
Bike purchase (choice)
People pick one offer
36 CBC tasks on e-bike price vs range — MNL on demo choices.
Steps
- Download the Bike purchase CBC kit.
- Review or edit choice_observations.csv.
- Run bundled from the sample shelf.
- Open Report for part-worths from MNL.
| File | Download |
|---|---|
attributes.csv |
attributes.csv |
choice_observations.csv |
choice_observations.csv |
choice_tasks.csv |
choice_tasks.csv |
settings.csv |
settings.csv |
Car purchase
People score each offer
MSRP vs fuel economy — vehicle configuration trade-offs with share simulation.
Steps
- Download the Car purchase kit.
- Set ratings in profiles.csv for each MSRP/MPG profile.
- Run from Examples or upload your CSVs.
- Review market share on Report.
| File | Download |
|---|---|
attributes.csv |
attributes.csv |
choice_set.csv |
choice_set.csv |
choice_tasks.csv |
choice_tasks.csv |
profiles.csv |
profiles.csv |
settings.csv |
settings.csv |
Car purchase short
People pick one offer
8 CBC tasks on MSRP vs fuel economy. A shorter sample than the 36-task car study.
Steps
- Download the short Car purchase kit.
- Eight tasks are enough to try the choice report.
- Run from the sample shelf.
- Review MNL results on the report.
| File | Download |
|---|---|
attributes.csv |
attributes.csv |
choice_observations.csv |
choice_observations.csv |
choice_tasks.csv |
choice_tasks.csv |
settings.csv |
settings.csv |
Car purchase (choice)
People pick one offer
CBC study on MSRP vs fuel economy — choice-based conjoint with MNL report.
Steps
- Download the Car purchase CBC zip.
- Edit demo choices or upload your respondent file.
- Run from the sample shelf.
- Review MNL results on Report.
| File | Download |
|---|---|
attributes.csv |
attributes.csv |
choice_observations.csv |
choice_observations.csv |
choice_tasks.csv |
choice_tasks.csv |
settings.csv |
settings.csv |
Column guide — what each file means
First-time users: you usually only edit rating values in
profiles.csv. Keep attribute level text identical to the sample files unless
you are designing a new study.
profiles.csv
required
One row per design profile you asked the respondent to rate. Each attribute column must use a level label exactly as listed in attributes.csv.
| Column | Meaning | Example |
|---|---|---|
(id column) |
Unique profile id — column name from settings id_key (e.g. name, profile, S1). One row per design profile. | S1 or 1 |
(attribute columns) |
One column per attribute; header = profile_key from attributes.csv. Cell value = one level label for that profile. | $1.55 or Flip |
rating |
Respondent score 0–100 (higher = more preferred). Replace demo values with your respondent's scores. | 85 |
rank |
Alternative to rating: rank order 1 = best, 2 = second, … Use when response_mode=rank in settings.csv. | 3 |
- Every design profile needs a rating (or rank) — missing rows cause errors.
- Level text must match attributes.csv exactly (spacing and $ signs matter for transport).
- In Excel: save as CSV UTF-8; avoid extra blank rows at the bottom.
attributes.csv
required
Names each feature, the levels a person can be offered, and the column name used in the other files.
| Column | Meaning | Example |
|---|---|---|
attribute |
Display name (report labels, RIMP). | Fare |
profile_key |
Column header used in profiles.csv and catalog/choice_set. | fare |
levels |
All level labels separated by | (pipe). | $1.55|$1.85|$2.15 |
baseline |
One of the levels. Leave it as the sample wrote it when you start from a sample. | $2.15 |
dummy_prefix |
Leave this cell as the sample wrote it. | XF |
level_format |
Optional display template for ordered attrs. {value} = numeric part. | {value} h |
value_type |
nominal (default), ordered, or ideal — affects prediction/interpolation. | ordered |
level_values |
For ordered attrs: numeric values parallel to levels (| separated). | 1.55|1.85|2.15 |
- baseline must be one of the levels.
- Ordered price/wait attributes enable interpolation between design levels.
choice_tasks.csv
optional
The questions. Each task is two or more offers, and each offer is one row. You do not write this file by hand. Attach attributes.csv and leave this file empty. The site writes the questions from the features and the levels, two offers in each.
| Column | Meaning | Example |
|---|---|---|
task_id |
The question. The same number on every offer in that question. | 1 |
alt_id |
The offer inside that question. | 1 |
(one column per feature) |
The column name is the profile_key from attributes.csv. The cell is one level of that feature. | $24000 |
- The number of questions follows the features and the levels. A file you supply is kept.
- Fewer than 12 answered questions still run. They are not steady enough to add to the shelf.
choice_observations.csv
optional
The picks. Who answered, and which offer they picked. A question with no offer selected is left out. On the study page, change any selected offer and choose Save & run. That writes this file.
| Column | Meaning | Example |
|---|---|---|
respondent_id |
The person. A new study uses R1. | R1 |
task_id |
The same question number as in choice_tasks.csv. | 1 |
alt_id |
The same offer number as in choice_tasks.csv. | 1 |
chosen |
1 on the offer they picked. 0 on the other offer in that task. | 1 |
(one column per feature) |
The same columns and the same level text as in choice_tasks.csv. | 40 mpg |
- A question with no offer selected is not in this file.
settings.csv
optional
Leave this file out and the study still runs. Two columns. The first row is key,value.
| Column | Meaning | Example |
|---|---|---|
key |
Setting name. | scenario_name |
value |
Setting value. | My car picks |
- scenario_name is the title on the result.
- id_key is the id column in profiles.csv, such as name. A pick-one file does not use id_key. Those rows are task_id and alt_id.
- holdout_ids lists offer ids left out of the fit, separated by commas, so the result can be checked against them. The Bike purchase sample uses S5.
- what_if is an extra offer that is not a row in profiles.csv. Write column=level, and separate features with |. A number between two listed levels is allowed on this row.
- tradeoff_reference is the offer the trade-off starts from. Use the same shape. Every value must be a listed level. From that offer, the full analysis text takes one step to a better level of one feature and says how much of the other feature that step is worth. This line does not change the report on this site, including If a price goes up. Leave the row out and the start is each feature's baseline. A value that is not a listed level leaves that trade-off out of the full text.
- response_mode is rating, or rank when 1 is best. The usual value is rating.
- rating_key is the score column when response_mode is rating. The usual name is rating. Set it only when the column has another name.
- rank_key is the rank column when response_mode is rank. The usual name is rank.
- interpolate_ordered stays on unless the value is false. While it is on, a level between two listed levels can be estimated. A sample that leaves this row out stays on.
- coding_scheme: leave this line. It is how the levels are compared, and the sample already chose it.
- prefill_tasks is how many questions start with an offer selected. The usual value is 12. A survey shows that many questions to each person. Leave the row out and the number is 12.
- A pick-one file does not use id_key, rating_key, rank_key, or holdout_ids.
catalog.csv
optional
Offers to place in order from the result.
| Column | Meaning | Example |
|---|---|---|
item_id |
Unique SKU id. | SKU01 |
name |
Display name in reports. | Alpha Flip Pro |
(attribute columns) |
Same profile_key columns as profiles.csv. | Nokia, Flip, 9, 115, 6 |
- Levels should match design levels when possible; off-design rows are flagged.
shortlist.csv
optional
Subset of catalog item_ids the respondent is considering.
| Column | Meaning | Example |
|---|---|---|
item_id |
One column listing catalog ids. | SKU01 |
- Report highlights best in shortlist vs best overall catalog.
choice_set.csv
optional
Offers compared on the result.
| Column | Meaning | Example |
|---|---|---|
option_id |
Unique option id. | A |
name |
Label in market share report. | Fast cheap |
(attribute columns) |
Same keys as attributes.csv. | $1.55, 10 min |
- Used for market share in the web report and run_share_simulation.py.
Worked examples (from sample kits)
Train commute — profiles rowname=S1, ticket_price=$1.50, platform_wait=8 min, rating=100 (your commute preference)
Coffee subscription — profiles rowname=S1, monthly_price=$12, delivery_days=Next day, rating=88 (your bundle preference)
Hotel weekend — profiles rowname=S1, nightly_rate=$89, distance_center=5 min walk, rating=72
Streaming plan — profiles rowname=S1, monthly_fee=$7.99, ads_per_hour=No ads, rating=91
Blueberry phones — profiles rowprofile=2, brand=Blueberry, style=Flip, talk_time_h=9, weight_g=115, camera_mp=2, rating=95