Allowed R6 · Within policy
{
"slug": "statistics-and-sql-screen",
"title": "Statistics and SQL Screen",
"difficulty": "medium",
"tags": [
"sql",
"data",
"technical",
"data-sci"
],
"is_free": true,
"locked": false,
"frequency": 0.25,
"last_reported_at": "2025-07-13T10:12:56.714334Z",
"url": "https://www.mokabla.com/en/c/huawei/statistics-and-sql-screen",
"agent": {
"page": "https://www.mokabla.com/agent/en/c/huawei/statistics-and-sql-screen",
"question": "https://www.mokabla.com/agent/_call/get_question?company=huawei&question=statistics-and-sql-screen"
},
"company": {
"slug": "huawei",
"name_en": "Huawei Egypt"
},
"body_md": "## What they ask\nFor data roles in the Cairo office the technical round leans almost entirely on two things: statistics fundamentals and SQL. The interviewer alternates between short conceptual questions and small query exercises you talk through out loud.\n\nQuestions in this area, for example:\n\n**Statistics**\n- When is the median a better summary than the mean, and give a telecom example.\n- Explain a p-value to a non-statistician without saying \"probability of the null\".\n- You have KPI samples from two groups of cell sites; which test would you use to compare them and what assumptions are you making?\n\n**SQL**, on a table `alarms(site_id, severity, opened_at, closed_at)`:\n1. Sites that raised more than five `critical` alarms last month.\n2. Average time to close, per severity, for alarms that were actually closed.\n3. For each site, the most recent alarm (one row per site).\n\n## Example\nFor item 3 the expected answer is a window function (`ROW_NUMBER() OVER (PARTITION BY site_id ORDER BY opened_at DESC)`) filtered to row 1; a `GROUP BY` with `MAX(opened_at)` alone loses the other columns, and they will point that out.\n\n## What they look for\n- Comfortable, unhurried explanations of basic statistics; they care that you can reason about data, not recite formulas.\n- Correct handling of `NULL` in `closed_at`, date arithmetic and grouping.\n- Whether you ask about the size of the table before proposing a query, which signals real-world experience.",
"follow_ups_md": "- How would you detect sites whose alarm rate is anomalous compared with their own history?\n- What changes in your queries if `alarms` has a billion rows?",
"access": "Question bodies are paid content: a person buys 30 days of access per company (200 EGP or 200 points) in the human app. Agents get titles, tags and difficulty; bodies only for free questions."
}