brand-sentiment-analysis
所属分类:自然语言处理
开发工具:CSS
文件大小:2116KB
下载次数:0
上传日期:2022-07-06 20:10:47
上 传 者:
sh-1993
说明: 品牌情绪分析,脚本利用Heartex平台从新闻中构建品牌情绪分析
(brand-sentiment-analysis,Scripts utilizing Heartex platform to build brand sentiment analysis from the news)
文件列表:
LICENSE (11357, 2019-07-24)
demo.png (182682, 2019-07-24)
requirements.txt (121, 2019-07-24)
src (0, 2019-07-24)
src\config.json (995, 2019-07-24)
src\heartex (0, 2019-07-24)
src\heartex\__init__.py (17, 2019-07-24)
src\heartex\add_sentiment_chart.py (2921, 2019-07-24)
src\heartex\api.py (2223, 2019-07-24)
src\heartex\create_filter_project.py (1013, 2019-07-24)
src\heartex\create_sentiment_project.py (886, 2019-07-24)
src\heartex\predict_and_filter.py (1800, 2019-07-24)
src\service.py (8423, 2019-07-24)
src\static (0, 2019-07-24)
src\static\data (0, 2019-07-24)
src\static\data\output.json (429367, 2019-07-24)
src\static\favicon.ico (21822, 2019-07-24)
src\static\js (0, 2019-07-24)
src\static\js\Chart.bundle.min.js (209055, 2019-07-24)
src\static\js\Chart.plots.js (3099, 2019-07-24)
src\static\js\DateFormat.js (4017, 2019-07-24)
src\static\js\jquery-3.3.1.min.js (86927, 2019-07-24)
src\static\js\vue.js (93675, 2019-07-24)
src\static\logs (0, 2019-07-24)
src\static\semantic-ui (0, 2019-07-24)
src\static\semantic-ui\semantic.css (828900, 2019-07-24)
src\static\semantic-ui\semantic.js (736574, 2019-07-24)
src\static\semantic-ui\semantic.min.css (628512, 2019-07-24)
src\static\semantic-ui\semantic.min.js (275730, 2019-07-24)
src\static\semantic-ui\themes (0, 2019-07-24)
src\static\semantic-ui\themes\default (0, 2019-07-24)
src\static\semantic-ui\themes\default\assets (0, 2019-07-24)
src\static\semantic-ui\themes\default\assets\fonts (0, 2019-07-24)
src\static\semantic-ui\themes\default\assets\fonts\brand-icons.eot (129648, 2019-07-24)
src\static\semantic-ui\themes\default\assets\fonts\brand-icons.svg (691865, 2019-07-24)
src\static\semantic-ui\themes\default\assets\fonts\brand-icons.ttf (129344, 2019-07-24)
src\static\semantic-ui\themes\default\assets\fonts\brand-icons.woff (87544, 2019-07-24)
... ...
# Brand Sentiment
A set of scripts that makes sentiment analysis of your brand
based on Google News and Twitter news streams. It utilizes Heartex
platform to create a custom neural network to do the study
specifically for your brand
[Tutorial](https://heartex.net/use-case/sentiment)
![](https://github.com/heartexlabs/brand-sentiment-analysis/raw/master/demo.png)
# Installation
> Important. To make it work you need to obtain **Heartex token**, to do so [signup here](https://go.heartex.net/business/signup/?ref=github). We give you a free account with 10k API requests (with above
link only!).
```sh
# install
python3 -m venv bsa-env
source bsa-env/bin/active
pip install -r requirements
```
```sh
# configure
export TOKEN=""
export BRAND=""
```
# Create Sentiment Model
```sh
# first we need to grab news data
python src/get_google_news.py --pages=10 --query=$BRAND --output=news.csv
```
```sh
# create project on heartex
python src/create_sentiment_project.py --token=$TOKEN --input=news.csv
# you will get project id, save it here
export SENTIMENT_PROJECT_ID=""
```
Open up `src/config.json` and put **$TOKEN** and **$SENTIMENT_PROJECT_ID** there
# Run
Execute ``` python3 service.py config.json```
# Add your own data
> [TBD]
# Advanced: Filter Results
In case your brand may appear in different contexts, for example, with
the name of one of your products (ex: Apple Watch), you may want to
filter those occurrences first.
To do that we will use another type of model which is called a tagger
model. It learns when you tag relevant occurrences.
```sh
PRODUCTS="Apple,iOS,iPadOS,watchOS,macOS,MacPro,Pro Display"
```
```sh
# create Heartex project to filter news that are only relevent to your brand name
# you will get back a link where you need to train a neural network a little bit to make it understand what is relevent to you
python src/create_filter_project.py --token=$TOKEN --input=news.csv --labels=$PRODUCTS
# set project here
export FILTER_PROJECT=""
```
Now you have what is called a smart filter, edit config.json and include it there. You will see smart filter buttons on the index page.
近期下载者:
相关文件:
收藏者: