LinkedIn network analysis with AI, step by step

Prompts 1 to 3 and 5 run on files from your archive. Prompt 4 needs LinkedAsk, a record of your activity connected to your assistant. For how a live connection works, see what a LinkedIn MCP is and what it can read.

LinkedIn InMaps is retired

LinkedIn announced InMaps on its blog in January 2011 as "an interactive visual representation of your professional universe". LinkedIn colored each map by the groups in your career, such as colleagues from one employer or classmates from one school, and you needed 50 connections and a profile that was 75 percent complete to get one.

As of October 2026, the old InMaps address forwards to a LinkedIn Help article titled "InMaps no longer supported", which no longer loads. LinkedIn offers no network map in its place.

Download the files

Four of the five prompts run on your LinkedIn data archive.

1. In LinkedIn, open Me, then Settings & Privacy, then Data privacy, and click Get a copy of your data.
2. Choose Download larger data archive and request it. LinkedIn says the larger download arrives by email within 24 hours and stays available for 72 hours.
3. Download the ZIP, start a new chat in the assistant you use, and attach the files that the prompt names.

Two files in the ZIP matter here:

• Connections.csv. Name, profile URL, email, company, position and connection date for each of your 1st-degree connections. A few lines of notes sit above the header row, and most email cells are blank.
• messages.csv. The date, sender, content and profile URLs of each message, archived conversations included.

The archive has no list of who viewed your profile and none of the reactions other people left on your posts. LinkedIn's download page says it does not provide Who's Viewed Your Profile data. What you get in a LinkedIn data export lists every file.

Five prompts for your network

Prompts 1 to 3 and 5 run on the archive. Prompt 4 runs on LinkedAsk, covered in the next section.

1. Your connections by company

This prompt reads the Company column of Connections.csv.

Here is my LinkedIn Connections.csv. Skip the notes lines above the header.
Count my connections by Company and show the top 15 as a bar chart.
Merge obvious duplicates such as "Acme" and "Acme Inc" and tell me which ones you merged.
List how many rows have no company.

Some rows have no company, and some have a connection date and nothing else. The last line of the prompt counts the rows with no company, so you know how much the chart leaves out.

2. Your connections by role

Position is the column here. Ask the assistant to choose groups that fit your network.

Using the same file, group my connections by role from the Position column.
Use about eight groups that fit my network, for example founders, recruiters, engineers, sales.
Show the groups as a bar chart and give two example titles for each group.

If the first set of groups looks wrong, name your own. Merge recruiter and talent partner if you treat them the same.

3. Network growth by month

The Connected On column gives each connection a date, which is enough for a monthly count.

Count new connections per month from the Connected On column.
Draw a column chart of new connections per month and a line chart of the running total.
Mark the three biggest months and list who I connected with in each.

A month reads low when people you connected with that month later left your network, because the file has no row for them. Charting LinkedIn connection growth by month explains this and shows the same count in a spreadsheet.

4. People who engage back

This needs your profile viewers and the people who reacted to or commented on your posts. The archive has neither, so the prompt runs on a record of your activity, covered in the next section.

For the last 90 days, list the people who viewed my profile, reacted to my posts or commented on them.
Group by person and rank the ones who appear in more than one of those three lists.
For each of the top ten, say whether we are connected.

5. Contacts who went quiet

Join messages.csv and Connections.csv by profile URL.

Using messages.csv and Connections.csv, find the date of my last message with each connection.
Match people by profile URL, never by name.
List connections I exchanged at least five messages with, where the last message is more than six months old.
Sort by how many messages we exchanged.

Example answer (fictional people, for illustration): "Nina Patel, 23 messages, last on March 4. Ravi Koh, 11 messages, last on January 19." To turn that list into messages, use the follow-up prompts.

Run the same prompts on your live LinkedIn history

LinkedAsk is a Chrome extension that records your LinkedIn activity as you use the site, and you can connect LinkedAsk to your own AI, such as Claude or ChatGPT, to ask about it in a chat.

With the connection added, prompt 4 runs without a file. The assistant reads your profile viewers and the people who engaged with your posts, then ranks the people who show up in more than one list. For prompt 3 it can also read LinkedIn's connection total on each day LinkedAsk read it. Ask for a weekly chart of the profiles you visited and it draws one, and it saves the chart to your LinkedAsk dashboard when you ask.

The Claude setup guide and the ChatGPT setup guide walk through adding the connection.

Pick by situation

• A one-time look at where your network sits: the archive and prompts 1 to 3, in one sitting once the file arrives.
• Reconnecting with people you lost touch with: prompt 5, then the follow-up prompts above. The full set is in the prompts for your LinkedIn network.
• Finding a way into one company: prompt 1, then the warm introduction prompts.
• Knowing who engages with you week to week: a record of your activity, connected to your assistant.