Stop losing your saved Instagram reels: two fixes

I exported my Instagram saved posts and reels. Below is a table (or raw JSON) with: date saved, author username, kind (reel or p = photo post), my collection name, and the post shortcode. I saved all of this and never revisit it.
 
IMPORTANT: you can only see usernames, collection names, dates and whether it is a reel. You can NOT see the captions or the video. Do not pretend you know what a specific reel says. Where you infer a topic from a username or collection name, label it "(guess)".
 
Do this, in order:
1. SNAPSHOT: total saves, how many are reels, date range, how many I saved in the last 30 days, and how many are in no collection.
2. CATEGORIES: sort everything into 6-10 categories (use my collection names first, then infer from usernames). Show each category with the count, the top 3 accounts in it, and a one-line read on why I seem to save that kind of content.
3. MY HABITS: the 5 accounts I save most. The month I saved the most. One honest observation about what I collect versus what I probably act on.
4. WATCH TONIGHT: pick the 10 saves most worth my time. Prefer: reels from my most-saved accounts, my newest saves, and categories where I have saved a lot but likely done little. For each give the link as https://www.instagram.com/reel/SHORTCODE/ (use /p/ if kind is p), one line on why, and how confident the pick is (high, medium, guess).
5. UNSAVE CANDIDATES: up to 15 saves I could drop (old, from accounts I saved only once, outside my main categories). List them; I will decide.
6. A 3-line weekly routine I can follow to actually go through the rest without it feeling like homework.
 
Keep it short, plain English, no filler. Use a table for step 2 and step 4.
 
My data:
[PASTE HERE]

#!/usr/bin/env python3
"""Flatten Instagram's 'Download your information' saved files into one CSV.
Usage: python3 flatten.py <export_folder> [out.csv]
Reads only the JSON files you point it at. Nothing is sent anywhere."""
import sys, json, csv, pathlib, datetime, re
 
def entries(node):
    """Yield every list of entry-dicts that contain 'string_map_data'."""
    if isinstance(node, dict):
        for v in node.values():
            yield from entries(v)
    elif isinstance(node, list):
        if node and all(isinstance(x, dict) for x in node) and any("string_map_data" in x for x in node):
            yield node
        else:
            for x in node:
                yield from entries(x)
 
def main():
    root = pathlib.Path(sys.argv[1])
    out = sys.argv[2] if len(sys.argv) > 2 else "saved.csv"
    files = sorted(p for p in root.rglob("*.json") if "saved" in p.name.lower())
    if not files:
        sys.exit("No saved*.json found. Look inside your_instagram_activity/ for a 'saved' folder.")
    rows, seen = [], set()
    for f in files:
        data = json.loads(f.read_text(encoding="utf-8"))
        for lst in entries(data):
            collection = ""
            for e in lst:
                smd = e.get("string_map_data", {})
                href = ts = val = ""
                for v in smd.values():
                    href = href or v.get("href", "")
                    ts = ts or v.get("timestamp", "")
                    val = val or v.get("value", "")
                if href:                          # a saved item
                    author = e.get("title") or val
                    m = re.search(r"instagram\.com/(reel|reels|p|tv)/([^/?#]+)", href)
                    kind = {"reels": "reel", "tv": "video"}.get(m.group(1), m.group(1)) if m else "other"
                    key = (href, collection)
                    if key in seen: continue
                    seen.add(key)
                    date = datetime.datetime.fromtimestamp(int(ts)).strftime("%Y-%m-%d") if ts else ""
                    rows.append([date, author, kind, collection or "(uncollected)", m.group(2) if m else href])
                elif val:                          # a collection header
                    collection = val
    filed = {r[4] for r in rows if r[3] != "(uncollected)"}
    rows = [r for r in rows if r[3] != "(uncollected)" or r[4] not in filed]
    rows.sort(reverse=True)
    with open(out, "w", newline="", encoding="utf-8") as fh:
        w = csv.writer(fh); w.writerow(["saved_on", "author", "kind", "collection", "shortcode"])
        w.writerows(rows)
    print(f"{len(rows)} saved items -> {out}")
 
main()

python3 flatten_saved.py ~/Downloads/your-unzipped-folder saved.csv