How to Export Test Scores to Excel, Google Sheets or Your Gradebook

To export test scores to Excel, you download the results as a CSV file and open it in Excel or Google Sheets. CSV is plain text in rows and columns, so almost every spreadsheet can open it and many school gradebooks can import it. In Snap AI the button is Export as CSV. There is no separate .xlsx download, and you don't need one.

The examples below use one class of 32 students who sat a 40-question test, but the steps work for any CSV of grades.

What is in the CSV file

Each row is one paper, meaning one student (a paper can have several pages). Columns always come in this order:

Question headers follow your answer key, so a test split into parts gets headers such as I.3 or II.2. Once you fix a flagged question in Review, it drops out of Notes.

• It's a .csv, not an .xlsx. Save it as an Excel workbook or Google Sheet before adding formulas; a CSV stores only values.
• It's UTF-8 with a byte-order mark (BOM). That marker tells Excel the encoding, so names such as Nguyễn, Müller, José or 王 display correctly.
• There is no class column, no student ID and no correct/incorrect flag. The answer key isn't in the file either.
• The file is named after the grading batch in the web app, so a batch called 8B Unit 3 quiz downloads as 8B Unit 3 quiz.csv.

Grade one class per batch and give it a clear name; with no class column in the file, that keeps classes apart.

How to export test scores to Excel

Download the CSV from Snap AI

1. Finish grading, then open Review on any paper with ungraded questions so the exported scores are final.
2. On the batch's Grading results page in the web app, select Export as CSV. The same button sits on the Batch statistics page.
3. In the Android app, Export as CSV opens your phone's share sheet; save the file to Google Drive or send it to your computer.

Not graded yet? Start with how to grade multiple-choice tests with your phone.

Open the CSV in Excel

1. Double-click the file, or use File > Open. On a computer that uses a decimal point (US, UK, Australia and similar), the columns land in place.
2. If everything lands in column A or the scores look odd, close without saving. Instead, open the workbook with your class list and use Data > From Text/CSV, which loads the scores on a new sheet there.
3. Choose the file and select Import. Set Delimiter to Comma and check that File Origin shows Unicode (UTF-8).
4. Select Load, or Transform Data if you need to fix decimals first (next section).
5. Save as an Excel Workbook (.xlsx) before you add formulas.

A Notes cell holds several lines when a paper has more than one ungraded question. If a row looks split in two after From Text/CSV, open Transform Data, select the gear icon next to Source and set Line breaks to Ignore quoted line breaks (see Microsoft's Text/CSV connector page). Menus are from Excel for Windows; Mac menus differ slightly.

Open the CSV in Google Sheets

1. Open the spreadsheet you want the scores in, or create a blank one.
2. Go to File > Import, select Upload and choose the CSV.
3. For the import location, pick Insert new sheets so your class list stays untouched.
4. Leave the separator on Detect automatically (or choose Comma) and select Import.

Google's import help page lists the other import options.

Decimal comma or decimal point: fix scores that import wrong

Snap AI writes scores with a decimal point (7.5) and separates columns with commas, the standard CSV convention. If your computer uses a decimal comma, as in much of Europe and Latin America and in Vietnam, Indonesia or Russia, you may get everything in one column, scores stuck as text, or 7.5 turned into a date such as 7 May.

• Excel: import with Data > From Text/CSV, set Delimiter to Comma and Data Type Detection to Do not detect data types, then select Transform Data. Otherwise Excel reads 7.5 with your regional format and may turn it into 75 or an error. Right-click the Score column, choose Change Type > Using Locale, pick Decimal Number and English (United States), then Close & Load. Excel then shows 7,5 in your own format. Microsoft's Power Query data types guide explains the option.
• Google Sheets: before importing, open File > Settings, set Locale to United States or United Kingdom and select Save settings. Import, then switch the locale back. The scores stay numbers and display with your decimal comma.
• Quick fix in Excel: if scores came in as text (not dates), select the Score column and use Find and Replace to change each full stop (.) to a comma (,).

With a decimal comma, Excel and Google Sheets also separate formula arguments with semicolons: =XLOOKUP(A2;Scores!A:A;Scores!C:C;"Not found"). This guide uses commas, so swap them if your spreadsheet rejects a formula.

Exam codes with a leading zero, such as 014, may show as 14 after import. To show three digits again, give the column the custom number format 000 (Excel: Format Cells > Custom; Google Sheets: Format > Number > Custom number format).

Match scores to your class list with XLOOKUP or VLOOKUP

Snap AI has no class roster: names are read from each student's handwriting, and rows aren't in class-list order. So don't paste the Score column straight into your register; look each student up by name.

Put the CSV data in the same workbook as your class list (in Excel, use Data > From Text/CSV or right-click the CSV's tab and choose Move or Copy), and name that tab Scores. On your class list sheet, put names in column A from A2, then type one of these in B2 and fill down:

• XLOOKUP (Excel for Microsoft 365, Excel 2021 or later, Google Sheets): =XLOOKUP(A2,Scores!A:A,Scores!C:C,"Not found"). See Microsoft's XLOOKUP reference for the optional arguments.
• VLOOKUP (older Excel): =IFERROR(VLOOKUP(A2,Scores!A:C,3,FALSE),"Not found"). The 3 is the third column of A:C (Score), and FALSE forces an exact match.
• Exam code too: =XLOOKUP(A2,Scores!A:A,Scores!B:B,"Not found") in column C, if you gave several versions.

Fix the names that don't match

• Not found usually means the spelling differs: Maria López against Maria Lopez, a nickname or an extra space. Correct the name on the Scores sheet and the result updates.
• Check for duplicates with =COUNTIF(Scores!A:A,A2) in a spare column. Anything above 1 means two papers carry that name (two students who share it, or one paper added twice), and the lookup returns only the first.
• Still not found? The student was probably absent, or the paper was never added.
• Prevent it by asking students to write their full name as it appears on your list. The free answer sheet generator prints a name line.

Convert the 10-point score to a percentage or another scale

Snap AI scores each paper out of 10 by default. If you set a different total before grading (anything from 0.01 to 1000), divide by that total instead of 10. Below, the score sits in B2 of your class list.

For letter grades, put the percentage as a number (=B2*10) in D2 and use =IFS(D2>=90,"A",D2>=80,"B",D2>=70,"C",D2>=60,"D",TRUE,"F") with your school's cut-offs. IFS works in Excel 2019 or later and in Google Sheets. Rows that say Not found show an error until you fix the name.

Copy scores into your school gradebook or LMS

Snap AI doesn't connect to gradebooks, markbooks or learning management systems, so the CSV is the bridge. Most gradebooks accept a CSV import or a pasted column of scores. Screens differ by system, so check its help pages; this pattern works almost everywhere:

1. If your gradebook can export a class list or import template, download it. It usually carries the student IDs the gradebook matches on, which Snap AI doesn't read.
2. Pull the Snap AI scores into that file with XLOOKUP, converted to the scale your gradebook expects.
3. Turn formulas into plain numbers: copy the column and paste it back with Paste Values in Excel or Paste special > Values only in Google Sheets.
4. Paste the column into the gradebook with your list in the same order, or save a CSV and import it: File > Save As > CSV UTF-8 (Comma delimited) in Excel (it saves only the sheet you are on), File > Download > Comma Separated Values (.csv) in Google Sheets.
5. Spot-check a few students, including the top and bottom scores, against Review.

The CSV is a snapshot: if you change points in Review after downloading, export again.

Simple item analysis in Excel: percent correct per question

The answer columns show which questions most students missed and which wrong option they chose. The CSV holds answers, not correct/incorrect flags, so you add the key yourself. In the example, students fill rows 2 to 33, and questions 1 to 40 sit in columns E to AR.

1. On the Scores sheet, type KEY in A35 and each correct answer under its question: E35 for question 1, F35 for question 2, and so on to AR35.
2. In E36, enter =COUNTIF(E$2:E$33,E$35)/ROWS(E$2:E$33), fill it right to AR36 and format the row as a percentage. Blank answers count as not correct.
3. Highlight weak questions, for example cells below 50%, with Home > Conditional Formatting in Excel or Format > Conditional formatting in Google Sheets.
4. For a weak question, count each option: =COUNTIF(E$2:E$33,"A"), then "B", "C" and "D". A popular wrong option points to an ambiguous question, a topic to reteach, or a wrong key.

With several versions, question 5 on code 101 isn't question 5 on code 102, so analyze each version separately. Instead of one key row, type one per code with the code in column B (B35 = 101, B36 = 102), and in E38 use =COUNTIFS($B$2:$B$33,$B$35,E$2:E$33,E$35)/COUNTIF($B$2:$B$33,$B$35). Copy it to E39, change each 35 to 36, then fill both rows right to AR. The multiple exam versions guide explains how Snap AI matches papers to keys.

The answer columns show what the AI read, not your corrections. If you changed points in Review, Score includes the change but the answer cell doesn't, so percent correct can differ slightly from the app. COUNTIF can also miss short numeric answers written differently, such as 12,5 against a key of 12.5, which Snap AI's scoring counts as the same number.

Or skip the spreadsheet: Snap AI's statistics page

For a quick look, select View statistics on a graded batch in the web app. Batch statistics shows the class average, an Answers by question chart counting correct, incorrect and ungraded answers for each question (one chart per exam code), and the lowest, average and highest score per exam code. It uses your reviewed results.

Use a spreadsheet for option-by-option counts or a term-long record. If one question has many ungraded answers, open those papers in Review; what affects AI reading accuracy covers the usual causes.

Keep student data tidy

A score export is personal data about your students, and you stay responsible for where it goes.

• Keep one master copy in your gradebook or class workbook, and name files with class, test and date.
• Delete files you no longer need: old exports in Downloads, copies emailed to yourself, files shared from your phone.
• Share individually. Send each student or parent their own score, never the whole class file.
• Follow your school's rules on where grade files may be stored.
• Know where photos go: papers you add to Snap AI are uploaded to cloud storage and read by an AI model (Google Gemini). Photograph only what you need for grading.