Image Virality

---

Comments-Caption Analysis Input -> comments.json, caption.json

---

• Clean: remove number, html code, punctation, capslock.

  • clean_data.py -> input: input_json/comments.json

• Compare: foreach post compare comments and caption to find word match.

  • compare.py -> input: input_json/comments.json -> output: output/comment-caption.json

• Find Similarity: foreach post cluster similar word(k-means), predict caption, dim reduction t-SNE and plot the results.

  • similarity.py -> input: output/comments.json

---