license: apache-2.0
task_categories:
- text-classification
- feature-extraction
language:
- ar
tags:
- original
- official
- real
- price
- freelance
- '2024'
pretty_name: khamsat.com
size_categories:
- 1K<n<10K
Khamsat
A structured Arabic-language dataset collected from khamsat.com, the largest Arabic freelance microservice marketplace.
This dataset is an independent research, not affiliated with khamsat. All content rights reserved to hsoub.com.
Abstract
Pricing in freelance marketplaces is a persistent challenge for both sellers seeking to maximize earnings and buyers seeking fair value. This dataset was constructed to support data-driven research into pricing dynamics on the Khamsat platform. It encompasses structured metadata, free-text service descriptions, keyword tags, and multi-dimensional buyer reviews.
Profile
Features
The dataset contains 29 fields grouped into four scopes: category, offer, owner, review
Click for more details
| Field | Type | Description |
|---|---|---|
category_name |
string | Top-level category |
category_url |
string | Category page URL |
service_name |
string | Sub-category |
service_url |
string | Sub-category URL |
offer_name |
string | Service title |
offer_url |
string | Page URL |
offer_stars |
float | Average rating |
offer_raters |
int | Number of raters |
offer_buyers |
int | Completed purchases |
offer_pending |
int | Active orders |
offer_price |
string | Price in USD |
offer_duration |
string | Delivery time |
offer_response_time |
string | Seller response time |
offer_available_additions |
string | Add-on options count |
offer_additions_price |
float | Add-on price |
offer_description |
string | Full service description |
offer_keywords |
list[string] | Associated tags |
offer_total_reviews |
int | Total buyer reviews |
offer_reviews |
list[dict] | Structured reviews |
owner_name |
string | Seller display name |
owner_url |
string | Seller profile URL |
owner_verified |
bool | Identity verified |
owner_level |
string | Platform tier |
owner_stars |
float | Overall rating |
owner_raters |
int | Number of raters |
owner_completion_rate |
string | Order completion rate |
owner_services |
int | Total services listed |
owner_customers |
int | Unique customers served |
owner_response_time |
string | Average response time |
Reviews
Each entry in offer_reviews contains the following fields.
Click for more details
| Field | Type | Description |
|---|---|---|
reviewer |
string | Buyer name |
rating_service |
float | Service quality / 5 |
rating_communication |
float | Communication / 5 |
rating_delivery |
float | On-time delivery / 5 |
date |
string | Review date |
comment |
string | Buyer comment |
Collection
Data was collected in two phases using Python-based web scraping tools.
The first phase used Selenium with Chrome WebDriver to navigate the Khamsat platform systematically across all service categories. Offer-level metadata and seller profile attributes were extracted for each listing, yielding 7,818 unique entries across 11 categories and 26 structured fields with no duplicates.
The second phase used Requests and BeautifulSoup to retrieve the textual content of each offer page individually. Three additional fields were extracted per offer: the free-text service description, keyword tags, and the full set of buyer reviews including per-dimension ratings. Requests were rate-limited with randomized delays between 1 and 2.5 seconds per thread across five concurrent workers to avoid overloading the platform.
Statistics
| Property | Value |
|---|---|
| Total offers | 7,818 |
| Service categories | 11 |
| Structured fields | 29 |
| Placeholder rate | 15.41% |
| Price range | $5 to $50 |
| Dominant price | $5 (54.8% of offers) |
| Dominant category | Programming and Development (19.3%) |
| Dominant delivery time | One day (49.6% of offers) |
Usage
from datasets import load_dataset
khamsat = load_dataset("IsmaelMousa/khamsat", split="train")
print(khamsat["offer_description"][0])
print(khamsat["offer_reviews"][0])
اهلا ومرحبا
سأصمم لكم فيديو احترافي رائع ومميز سواءا لصفحتك على الفيسبوك او لموقعك الإلكتروني او لقناتك على اليوتيوب او لحملة اعلانية لشركاتك او للترويج لنشاطك التجاري أو بكل بساطة ألبوم صور على شكل فيديو يوثق ذكرياتك اوعيد ميلاد او زفاف او حفلة تخرج...إلخ.
سعر الخدمة يتضمن مونتاج 60 ثانية مونتاج جيد مع انتقالات جيدة ولا يشمل الاعلانات .اما ادا كنت تريد امكانيات عالية وفاخرة على مستوى الصوت والصورة فيرجى الاطلاع على التطويرات.
*العمل لا يسلم بصوره مفتوحة المصدر
*جودة الفيديو full hd
كن على يقين انك لن ترحل إلا وانت راضي تماما عن الخدمة .
[{'reviewer': 'بشير عمر ب.', 'rating_service': 5.0, 'rating_communication': 5.0, 'rating_delivery': 5.0, 'date': 'منذ 5 أيام و12 ساعة', 'comment': 'خمس نجوم، إنه مونتاج احترافي جداً، وأنصح الجميع بالتعامل معه'}, {'reviewer': 'وسيم ا.', 'rating_service': 5.0, 'rating_communication': 5.0, 'rating_delivery': 5.0, 'date': 'منذ 11 يوم و4 ساعات', 'comment': 'تجربة أكثر من رائعة! أستاذ يونس فنان ومبدع حقيقي وليس مجرد منفذ. استلم مشروع فيديو إعلاني لمنتج عالي القيمة (High-Ticket) بستايل سينمائي فخم (Dark Mode و 3D)، والنتيجة كانت احترافية وفوق المتوقع بكثير.\n\nيتميز بمهارة عالية في تحريك النصوص (Typography) وتعديل الألوان بدقة. شخص راقي جداً في التواصل، صبور، ومتجاوب مع الملاحظات والتعديلات برحابة صدر. إذا كنت تبحث عن جودة (VIP) ومستوى إخراج فاخر لمشروعك، يونس هو الخيار الأمثل. كل الشكر لك على هذا الإبداع، وأكيد رح يكون بينا تعاون مستمر بإذن الله.'}]
Citation
If you use this dataset or reference the companion study, please cite the following.
@misc{mousa2024khamsat,
author = {Ismael Mousa},
title = {Data-Driven Optimization of Pricing Strategies on Microservice Platforms: Insights from Khamsat},
year = {2024},
note = {https://github.com/IsmaelMousa/khamsat-predictor}
}
License
This dataset is released under the Apache 2.0 License. All original content scraped from khamsat.com remains the intellectual property of Hsoub and its contributors.