Summary
In this video [0:04], Landon from RH explores using Google Gemini for survey exam prep, specifically for FEMA flood map interpretation. Despite an initial setback with Gemini’s inability to directly read a FEMA firm map [0:21], Landon cleverly encodes the map data into a text file [0:28] and combines it with sample elevation certificate data and the FEMA elevation certificate PDF form [0:42]. After uploading these files to Gemini and providing specific instructions, the AI generates a range of sample exam questions [1:14], including multiple-choice, true/false, and even word problems [2:40]. While the results aren’t perfect and may require some tweaking [3:03], Landon concludes that this approach is a significant improvement over starting from scratch for creating survey exam questions [3:10], even if Gemini can’t quite replace human expertise just yet.
FAQ
#:1 What problem was Landon trying to solve in this video?
Landon was exploring how to use Google Gemini to help with survey exam preparation, specifically focusing on the ability to interpret FEMA flood maps and understand elevation certificates [0:04]. The challenge was to see if Gemini could generate relevant test questions from a flood map and related documents, to evaluate someone’s ability to read the map and fill out an elevation certificate correctly.
#:2 What was the initial issue Landon encountered with Google Gemini?
Initially, Google Gemini was unable to directly read the FEMA firm map that Landon uploaded [0:21]. The AI couldn’t extract the necessary text or data from the raster image of the flood map, which prevented it from generating relevant test questions based solely on the map image.
#3: How did Landon work around Gemini’s inability to read the flood map directly?
To overcome this limitation, Landon extracted the data from the flood map and encoded it into a text file [0:28]. He then uploaded this text file, along with sample elevation certificate data and the FEMA elevation certificate PDF form, to Google Gemini. By providing the map data in a text format, Gemini was able to understand and use the information to generate sample exam questions [0:42].
#4: What types of exam questions was Gemini able to create?
After uploading the files and providing instructions, Google Gemini generated various types of sample exam questions, including multiple-choice questions, true/false statements, and even word problems [1:14]. These questions were designed to test the user’s ability to read and understand the flood map data, as well as their knowledge of how to correctly fill out an elevation certificate form. While the results weren’t perfect and required some tweaking, Landon found the AI-generated questions to be a good starting point for exam preparation [3:03].