Using UIAutomator for Mobile App Data Scraping and Quality Evaluation in K12 Education Apps

Introduction: In mobile product competitive evaluation, capturing data from competitor apps and assessing content quality often encounters technical obstacles; this article shares several practical solutions.

Task background: The quality assessment of search results in K12 online‑education apps follows the workflow illustrated in the diagram below.

Common approaches and drawbacks: Using packet‑capture tools to discover API URLs and manually constructing HTTP requests is limited because many apps enforce hardware token validation, use long‑living socket connections, encrypt responses, or accept only image inputs, making standard HTTP scraping ineffective.

Therefore, an automation‑driven solution is required, addressing three key problems:

Automation driver: UIAutomator is chosen because it supports cross‑application automation, operates on control IDs, and offers a good balance of cost, usability, and efficiency.

Non‑standard image input: Since the K12 app’s input is an image and each competitor crops images differently, a normalization step is introduced. By combining the screenshot with a background canvas and applying the target app’s cropping strategy, consistent image regions are obtained (see the normalization diagrams below).

Structured UI control reverse‑engineering: UIAutomator’s UI‑hierarchy parsing extracts control attributes, allowing both text‑based and image‑based data collection without relying on network requests.

Additional common issues and solutions:

App usage limits: Apps often track daily usage via a device ID derived from IMEI and MAC address. By using an MTK‑based Android device with Xposed modules, these identifiers can be altered within the UIAutomator script.

Low automation efficiency: Large‑scale batch analysis requires multiple machines; a resource‑for‑time trade‑off strategy is adopted.

Conclusion: The case study demonstrates that UIAutomator can automate data collection and quality evaluation for competitor K12 education apps, effectively overcoming device identity protection, non‑standard image inputs, and UI encryption through device spoofing, image normalization, and UI control reverse‑engineering.

Author: Yin Fei, Senior Test Engineer at Baidu Nuomi, with extensive experience in mobile testing tools.