john-adeojo

import transformers
from transformers import AutoModelForSequenceClassification, AutoTokenizer, pipeline
from scripts.text_processing.preprocess_tweets_lite import TextCleaner
import pandas as pd

class SentimentAnalyzer(TextCleaner):
    def __init__(self, model="cardiffnlp/twitter-roberta-base-sentiment-latest", emotion=False):
        super().__init__(stop_words_remove=False)
        self.model = None
        self.tokenizer = None

john-adeojo

# calculate per follower metrics 
tweets_df['favorite_count_pf'] = tweets_df['favorite_count']/ tweets_df['follower_count']
tweets_df['retweet_count_pf'] = tweets_df['retweet_count']/ tweets_df['follower_count']
tweets_df['quote_count_pf'] = tweets_df['quote_count']/ tweets_df['follower_count']
tweets_df['reply_count_pf'] = tweets_df['reply_count']/ tweets_df['follower_count']

# normalise metrics
scaler = MinMaxScaler()
scaled_values = scaler.fit_transform(tweets_df[['favorite_count_pf', 'retweet_count_pf', 'quote_count_pf', 'reply_count_pf']])
tweets_df[['favorite_count_pf', 'retweet_count_pf', 'quote_count_pf', 'reply_count_pf']] = scaled_values

john-adeojo

from sklearn.linear_model import LinearRegression

# Create a list of the tuned models with names
models = [
    ('xgb', xgb_tuner.best_estimator_),
    ('rf', rf_tuner.best_estimator_),
    ('ridge', ridge_tuner.best_estimator_)
]

# Create the stacking model

john-adeojo

from skopt import BayesSearchCV

# Tune XGBoost model
xgb_tuner = BayesSearchCV(
    xgb_pipeline,
    xgb_param_grid,
    cv=5,
    scoring='neg_root_mean_squared_error',
    n_iter=30,
    n_jobs=-1

john-adeojo

from skopt.space import Real, Integer, Categorical

# Parameter grid for XGBoost
xgb_param_grid = {
    'regressor__learning_rate': Real(0.01, 0.3, prior='log-uniform'),
    'regressor__n_estimators': Integer(50, 2000),
    'regressor__max_depth': Integer(3, 50),
    'regressor__min_child_weight': Integer(1, 20),
    'regressor__gamma': Real(0, 5),
    'regressor__subsample': Real(0.5, 1),

john-adeojo

from sklearn.compose import ColumnTransformer
from sklearn.pipeline import Pipeline
from sklearn.ensemble import RandomForestRegressor, StackingRegressor
from sklearn.linear_model import Ridge, LinearRegression
from sklearn.preprocessing import OneHotEncoder, MinMaxScaler, StandardScaler
from sklearn.model_selection import train_test_split
from xgboost import XGBRegressor


preprocessor = ColumnTransformer(

john-adeojo

# define funtion te rplace missing vairbales

import pandas as pd

def impute_missing(df, test=False):
    
    if test == False:
        id_df = df['Id']
        y = df['SalePrice']
        df = df.drop(columns=['Id', 'SalePrice'])

john-adeojo

import streamlit as st
import pandas as pd
import folium
from folium.plugins import MarkerCluster
from streamlit_folium import folium_static

tweets_dash_final =pd.read_csv(path_to_csv)

# Define a function to assign emojis based on emotion
def get_emoji(sentiment):

john-adeojo

import re
from nltk import word_tokenize, WordNetLemmatizer
from nltk.corpus import stopwords


class TextCleaner:
    def __init__(self, stop_words=None, stop_words_remove=False):
        
        self.stop_words_remove = stop_words_remove
        if stop_words:

john-adeojo

import pandas as pd
import torch
from sklearn.model_selection import train_test_split
from torch.utils.data import Dataset, DataLoader
from transformers import AutoTokenizer


class DataPipeline:
    def __init__(self, df, target_col, text, model, save_data, random_state=42):
        self.df = df