![\"\"](\"https:\/\/miro.medium.com\/v2\/resize:fit:700\/1*S5rzIPoLJfGNP6bZet0Seg.png\")
<\/figure>Time series analysis refers to techniques for deriving useful statistics and other data properties from time series data. Time series forecasting is the process of using a model to forecast future values based on values that have already been observed.<\/p>\n
Time series analysis is used in our day-to-day activities as it answers the questions of \u201cHow many\u201d, \u201cWhen will\u201d, \u201cHow will\u201d, \u201cHow much\u201d and other related questions that may arise when there is a need to predict the future. Depending on the level of accuracy required for the prediction, the duration of the predicted period, and, of course, the amount of time available for feature selection and parameter tuning to produce the desired results, we can approach prediction tasks in various ways.<\/p>\n
Non-stationary data such as economics, meteorology, stock prices, and retail sales are presented using time series, but must usually be made stationary during pre-processing for predictions. With Prophet, however, we can skip this step and feed non-stationary data directly to the model. We\u2019ll demonstrate how to do this in this article.<\/p>\n Comet is a platform for experimentation that enables you to monitor your machine-learning experiments. Comet has another noteworthy feature: it allows us to conduct exploratory data analysis. We can accomplish our EDA objectives thanks to Comet\u2019s integration with well-known Python visualization frameworks. You can learn more about Comet here<\/a>.<\/p>\n This lesson will teach us how to integrate Comet with our time-series forecasting model. We will carry out some EDA on our Tesla stock dataset, and then we will log the visualizations onto the Comet experimentation website or platform. Without further ado, let\u2019s begin.<\/p>\n You may install the Comet library on your computer if you don\u2019t already have it there by using either of the following lines at the command prompt. Note that if you are installing packages directly into a Colab notebook, or any environment that uses virtual machines, you\u2019ll likely want to use \u2014 or \u2014<\/p>\n Prophet is a additive model for predicting time series data that fits non-linear trends with seasonality that occurs annually, weekly, daily, and on weekends as well as during holidays. Strongly seasonal time series and multiple seasons of historical data are ideal for it. Prophet typically manages outliers well and is robust to missing data and changes in trends.<\/p>\n The Facebook Core Data Science team recently released Prophet as an open-source software. It can be downloaded from PyPI and CRAN. To learn more about Prophet, you can check it out here<\/a>.<\/p>\n Before we begin our project, we will import Comet and Prophet using the code below.<\/p>\n We then create our Next, we can get started on our car monthly sales time analysis project. We begin by importing all the remaining necessary libraries.<\/p>\n After importing all the necessary libraries, we now import our dataset<\/strong><\/a>.<\/p>\n We then show the first 10 lines of our dataset using the code below:<\/p>\n In order to use the Before we fit our model, let\u2019s quickly visualize it first:<\/p>\n We then fit our model:<\/p>\n We then make a forecast to predict the price of cars in the future, where N is the number of time periods we would like to predict. Because our data is monthly, this means we will predict the next 10 months worth of data:<\/p>\n We can now plot the prediction:<\/p>\n From the forecast, we can see the rise and fall of prices in cars over the years, with car prices reaching their peak in 1968 at 26,000 before dropping off the ladder in 1969.<\/p>\n<\/div>\n<\/div>\n<\/div>\n\n\n\n Curious to see how Comet works? Check out our PetCam scenario to see MLOps in action.<\/a><\/p><\/blockquote>\n<\/div>\n<\/div>\n<\/div>\n\n\n\n The next step involves viewing the components of our time series analysis which includes sales per year and other features.<\/p>\n Finally, we must view our results in Comet, which requires us to terminate our experiment.<\/p>\n To check your result, follow the link displayed in your notebook when you started the experiment, or login to your Comet account and head over to your workspace.<\/p>\n The tuned parameters are listed in the hyperparameters section:<\/p>\n The logged model can be found in the Assets & Artifacts menu.<\/p>\n Under the Graphics menu, you can see the produced We just tracked our monthly car price prediction model in Comet.<\/p>\n In this tutorial, we\u2019ve demonstrated how to forecast time series data in Comet using Prophet. We\u2019ve discussed how to set up your data, make predictions, and monitor the outcomes in Comet.<\/p>\n To work with the dataset, you can get it here<\/a>. You can also download the full code used in this tutorial here<\/a>.<\/p>\n<\/div>\n<\/div>\n<\/div>\n","protected":false},"excerpt":{"rendered":" Photo credit: z5.ai Time series analysis refers to techniques for deriving useful statistics and other data properties from time series data. Time series forecasting is the process of using a model to forecast future values based on values that have already been observed. Time series analysis is used in our day-to-day activities as it answers […]<\/p>\n","protected":false},"author":99,"featured_media":0,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"customer_name":"","customer_description":"","customer_industry":"","customer_technologies":"","customer_logo":"","footnotes":""},"categories":[9,7],"tags":[],"coauthors":[197],"class_list":["post-7790","post","type-post","status-publish","format-standard","hentry","category-product","category-tutorials"],"yoast_head":"\n![\"A](\"https:\/\/miro.medium.com\/v2\/resize:fit:700\/1*JA-q7gCmpe7jrGcFOl4z3A.png\")
<\/figure>Comet<\/strong><\/h1>\n
Prerequisites<\/strong><\/h2>\n
pip<\/code>.<\/p>\npip install comet_ml<\/span><\/pre>\nconda install -c comet_ml<\/span><\/pre>\nProphet<\/strong><\/h1>\n
import comet_ml\nfrom prophet import Prophet<\/span><\/pre>\n.comet.config<\/code> file and and add in our API key, workspace, and name of our project so all the readings on our models can be automatically tracked and logged. If you don\u2019t already have a Comet account, you can sign up for free here<\/a>, and then just grab your API key from Account settings<\/code> \/ API Keys<\/code>.<\/p>\ncomet_string = \"\"\"[comet]\napi_key= <Your-API-Key>\nproject_name= \"Stock Price prediction\"\nworkspace= olujerry\n\"\"\"\nwith open('.comet.config', 'w') as f:\nf.write(comet_string)<\/span>exp = Experiment()<\/span><\/pre>\nimport numpy as np\nimport pandas as pd\nimport matplotlib.pyplot as plt\nfrom pandas.plotting import lag_plot\nfrom pandas import datetime\nfrom statsmodels.tsa.arima_model import ARIMA\nfrom sklearn.metrics import mean_squared_error<\/span><\/pre>\nCar_sales = pd.read_csv('\/content\/monthly-car-sales.csv',\n parse_dates=['Month'])<\/span><\/pre>\nCar_sales.head(10)<\/span><\/pre>\n![\"\"](\"https:\/\/miro.medium.com\/v2\/resize:fit:337\/1*Xf9DO_SKdctFFF0x4bmOBw.png\")
<\/figure>prophet.fit()<\/code> function, the DataFrame must be in a particular format. Datetime information must be in the first column, which must be named ds<\/code>. The observations must be in the second column, which must be named y<\/code>.<\/p>\nCar_sales= Car_sales.rename(columns= {'Month':'ds', 'Sales':'y'})<\/span><\/pre>\nplt.plot(Car_sales['ds'], Car_sales['y'])\nplt.show()<\/span><\/pre>\n![\"\"](\"https:\/\/miro.medium.com\/v2\/resize:fit:700\/1*j9GFHeW72hKjUsKX8XGpwA.png\")
<\/figure>Feature = Prophet()\nFeature.fit(Car_sales)<\/span><\/pre>\nN = 10\nfuture_Car_sales = Feature.make_future_dataframe(periods=N)\nforecast = Feature.predict(future_Car_sales)<\/span><\/pre>\nFeature.plot(forecast)\nplt.show()<\/span><\/pre>\n![\"\"](\"https:\/\/miro.medium.com\/v2\/resize:fit:700\/1*HE82Pagrr8mhil13sRJn4Q.png\")
<\/figure>Feature.plot_components(forecast)\nplt.show()<\/span><\/pre>\nexp.end()<\/span><\/pre>\nShowing results in Comet<\/h1>\n
![\"\"](\"https:\/\/miro.medium.com\/v2\/resize:fit:700\/1*_zC6-tEJawjFD0WNnjE-fw.png\")
<\/figure>![\"\"](\"https:\/\/miro.medium.com\/v2\/resize:fit:700\/1*NXDkJGZxmbgamX9ky_a2GA.png\")
<\/figure>![\"\"](\"https:\/\/miro.medium.com\/v2\/resize:fit:700\/1*M_AYY6Vg3FzB11h_9Agj1A.png\")
<\/figure>matplotlib<\/code> figures. This shows all the charts for every plot made in our notebook.<\/p>\n![\"\"](\"https:\/\/miro.medium.com\/v2\/resize:fit:700\/1*zWbBwpq1yfL66zK1jxAeeQ.png\")
<\/figure>Conclusion<\/h1>\n