{"id":1688432,"date":"2022-08-25T12:00:19","date_gmt":"2022-08-25T19:00:19","guid":{"rendered":"https:\/\/www.esri.com\/arcgis-blog\/?post_type=blog&p=1688432"},"modified":"2024-01-18T11:12:03","modified_gmt":"2024-01-18T19:12:03","slug":"automl-in-arcgis-pro-3-0","status":"publish","type":"blog","link":"https:\/\/www.esri.com\/arcgis-blog\/products\/arcgis-pro\/analytics\/automl-in-arcgis-pro-3-0","title":{"rendered":"AutoML in ArcGIS Pro 3.0"},"author":154341,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"open","ping_status":"closed","template":"","format":"standard","meta":{"_acf_changed":false,"_searchwp_excluded":""},"categories":[23341],"tags":[766982,665211,35661,759592],"industry":[],"product":[36561],"class_list":["post-1688432","blog","type-blog","status-publish","format-standard","hentry","category-analytics","tag-automl","tag-geoai","tag-machine-learning","tag-spatial-data-science","product-arcgis-pro"],"acf":{"authors":[{"ID":154341,"user_firstname":"Nicholas","user_lastname":"Giner","nickname":"Nick Giner","user_nicename":"nginer","display_name":"Nicholas Giner","user_email":"NGiner@esri.com","user_url":"","user_registered":"2021-01-07 14:31:25","user_description":"Nick Giner is a Product Manager for Spatial Analysis and Data Science. Prior to joining Esri in 2014, he completed Bachelor\u2019s and PhD degrees in Geography from Penn State University and Clark University, respectively. In his spare time, he likes to play guitar, golf, cook, cut the grass, and read\/watch shows about history.","user_avatar":""}],"short_description":"Learn about AutoML in the new GeoAI toolbox in ArcGIS Pro 3.0","flexible_content":[{"acf_fc_layout":"content","content":"

At ArcGIS Pro 3.0, we introduced the GeoAI toolbox<\/a>, which brings new AI and machine learning algorithms to ArcGIS Pro.\u00a0 This toolbox contains eight geoprocessing tools that help to solve Natural Language Processing problems (Text Analysis Toolset \u2013 6 tools), and classification and regression problems (Feature and Tabular Analysis Toolset \u2013 2 tools).<\/p>\n

In this blog article, we\u2019ll focus on the two tools within the Feature and Tabular Analysis toolset<\/a>:\u00a0 Train Using AutoML<\/a> and Predict Using AutoML<\/a>.\u00a0 We\u2019ll briefly explain what AutoML is and why it’s useful, and then see an example of it applied to a real-world spatial data science<\/a> problem.<\/p>\n"},{"acf_fc_layout":"image","image":{"ID":1689502,"id":1689502,"title":"toolbox_v2","filename":"toolbox_v2.png","filesize":49676,"url":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/toolbox_v2.png","link":"https:\/\/www.esri.com\/arcgis-blog\/products\/arcgis-pro\/analytics\/automl-in-arcgis-pro-3-0\/toolbox_v2","alt":"","author":"154341","description":"","caption":"","name":"toolbox_v2","status":"inherit","uploaded_to":1688432,"date":"2022-08-18 17:23:59","modified":"2022-08-18 17:23:59","menu_order":0,"mime_type":"image\/png","type":"image","subtype":"png","icon":"https:\/\/www.esri.com\/arcgis-blog\/wp-includes\/images\/media\/default.png","width":671,"height":655,"sizes":{"thumbnail":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/toolbox_v2-213x200.png","thumbnail-width":213,"thumbnail-height":200,"medium":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/toolbox_v2.png","medium-width":267,"medium-height":261,"medium_large":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/toolbox_v2.png","medium_large-width":671,"medium_large-height":655,"large":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/toolbox_v2.png","large-width":671,"large-height":655,"1536x1536":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/toolbox_v2.png","1536x1536-width":671,"1536x1536-height":655,"2048x2048":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/toolbox_v2.png","2048x2048-width":671,"2048x2048-height":655,"card_image":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/toolbox_v2-476x465.png","card_image-width":476,"card_image-height":465,"wide_image":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/toolbox_v2.png","wide_image-width":671,"wide_image-height":655}},"image_position":"center","orientation":"horizontal","hyperlink":""},{"acf_fc_layout":"content","content":"

What is AutoML, and why use it?<\/strong><\/h2>\n

ArcGIS has a powerful set of existing tools<\/a> that allow you to apply machine learning algorithms to solve complex, real-world problems.\u00a0 These tools, however, are typically not used in a vacuum; they are most often used as part of a broader set of steps within the machine learning process<\/a>.\u00a0 This process starts with identifying the business problem and formulating the problem statement or question, and is then followed by data collection and preprocessing, algorithm selection, model training and tuning, model evaluation, and finally prediction.<\/p>\n

 <\/p>\n"},{"acf_fc_layout":"image","image":{"ID":1688452,"id":1688452,"title":"ML_process","filename":"ML_process.png","filesize":9977,"url":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/ML_process.png","link":"https:\/\/www.esri.com\/arcgis-blog\/products\/arcgis-pro\/analytics\/automl-in-arcgis-pro-3-0\/ml_process","alt":"","author":"154341","description":"","caption":"Steps of the machine learning process.","name":"ml_process","status":"inherit","uploaded_to":1688432,"date":"2022-08-17 20:01:37","modified":"2022-09-01 15:25:11","menu_order":0,"mime_type":"image\/png","type":"image","subtype":"png","icon":"https:\/\/www.esri.com\/arcgis-blog\/wp-includes\/images\/media\/default.png","width":1110,"height":420,"sizes":{"thumbnail":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/ML_process-213x200.png","thumbnail-width":213,"thumbnail-height":200,"medium":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/ML_process.png","medium-width":464,"medium-height":176,"medium_large":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/ML_process.png","medium_large-width":768,"medium_large-height":291,"large":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/ML_process.png","large-width":1110,"large-height":420,"1536x1536":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/ML_process.png","1536x1536-width":1110,"1536x1536-height":420,"2048x2048":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/ML_process.png","2048x2048-width":1110,"2048x2048-height":420,"card_image":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/ML_process-826x313.png","card_image-width":826,"card_image-height":313,"wide_image":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/ML_process.png","wide_image-width":1110,"wide_image-height":420}},"image_position":"center","orientation":"horizontal","hyperlink":""},{"acf_fc_layout":"content","content":"

It is important to note that this is an iterative process that can involve extensive experimentation and repetition.\u00a0 You may get to the model training step, then realize that you need to go back a few steps to collect more data, preprocess it in a different way, or choose an alternative algorithm, then potentially do it all over again!<\/p>\n

Automated Machine Learning (AutoML) is an approach aimed to automate much of this process, including data preprocessing and engineering, model training, hyperparameter tuning, and model evaluation.\u00a0 In doing so, it allows the user to repeat this process for many different machine learning algorithms, which helps them identify the best algorithm for their specific problem and data.<\/p>\n"},{"acf_fc_layout":"image","image":{"ID":1688462,"id":1688462,"title":"AutoML_process","filename":"AutoML_process.png","filesize":11812,"url":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/AutoML_process.png","link":"https:\/\/www.esri.com\/arcgis-blog\/products\/arcgis-pro\/analytics\/automl-in-arcgis-pro-3-0\/automl_process","alt":"","author":"154341","description":"","caption":"Steps of the machine learning process. The grey box indicates steps that can be automated by AutoML.","name":"automl_process","status":"inherit","uploaded_to":1688432,"date":"2022-08-17 20:02:20","modified":"2022-09-01 15:25:04","menu_order":0,"mime_type":"image\/png","type":"image","subtype":"png","icon":"https:\/\/www.esri.com\/arcgis-blog\/wp-includes\/images\/media\/default.png","width":1110,"height":422,"sizes":{"thumbnail":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/AutoML_process-213x200.png","thumbnail-width":213,"thumbnail-height":200,"medium":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/AutoML_process.png","medium-width":464,"medium-height":176,"medium_large":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/AutoML_process.png","medium_large-width":768,"medium_large-height":292,"large":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/AutoML_process.png","large-width":1110,"large-height":422,"1536x1536":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/AutoML_process.png","1536x1536-width":1110,"1536x1536-height":422,"2048x2048":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/AutoML_process.png","2048x2048-width":1110,"2048x2048-height":422,"card_image":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/AutoML_process-826x314.png","card_image-width":826,"card_image-height":314,"wide_image":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/AutoML_process.png","wide_image-width":1110,"wide_image-height":422}},"image_position":"center","orientation":"horizontal","hyperlink":""},{"acf_fc_layout":"content","content":"

As a result of its automated nature, AutoML can be beneficial for users with all levels of machine learning experience.<\/p>\n

If you are new to machine learning, AutoML allows you to bring your data in as is, and quickly develop quality models that you can build upon.\u00a0 It also produces output reports with charts and explanations to help you understand some of the details that go on behind the scenes.<\/p>\n

If you already have a solid background and understanding of machine learning, AutoML can be used to do some of the time consuming iterating and experimenting tasks, and therefore you can focus your time and energy on other tasks that require your expertise.\u00a0 Think of AutoML here as a productivity booster, and as a way generate good baseline models with minimal human input and time.<\/p>\n

Using AutoML to study water access in Uganda<\/strong><\/h2>\n

To demonstrate the AutoML tools, I\u2019ll focus on the issue of clean water access in developing countries.<\/p>\n

In Uganda, nearly 90% of the population lives in small towns and rural areas, and roughly two thirds of this group lack access to safe water.\u00a0 Those that do have access often have to walk long distances to collect clean water from various types of pumps or tapstands (a.k.a. “water points”).<\/p>\n"},{"acf_fc_layout":"image","image":{"ID":1688982,"id":1688982,"title":"pump_CC","filename":"pump_CC.jpg","filesize":190563,"url":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/pump_CC.jpg","link":"https:\/\/www.esri.com\/arcgis-blog\/products\/arcgis-pro\/analytics\/automl-in-arcgis-pro-3-0\/pump_cc","alt":"","author":"154341","description":"","caption":"Source: Creative Commons (Julien Harneis).","name":"pump_cc","status":"inherit","uploaded_to":1688432,"date":"2022-08-18 12:43:16","modified":"2022-09-01 15:22:21","menu_order":0,"mime_type":"image\/jpeg","type":"image","subtype":"jpeg","icon":"https:\/\/www.esri.com\/arcgis-blog\/wp-includes\/images\/media\/default.png","width":1024,"height":681,"sizes":{"thumbnail":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/pump_CC-213x200.jpg","thumbnail-width":213,"thumbnail-height":200,"medium":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/pump_CC.jpg","medium-width":392,"medium-height":261,"medium_large":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/pump_CC.jpg","medium_large-width":768,"medium_large-height":511,"large":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/pump_CC.jpg","large-width":1024,"large-height":681,"1536x1536":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/pump_CC.jpg","1536x1536-width":1024,"1536x1536-height":681,"2048x2048":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/pump_CC.jpg","2048x2048-width":1024,"2048x2048-height":681,"card_image":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/pump_CC-699x465.jpg","card_image-width":699,"card_image-height":465,"wide_image":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/pump_CC.jpg","wide_image-width":1024,"wide_image-height":681}},"image_position":"center","orientation":"horizontal","hyperlink":""},{"acf_fc_layout":"content","content":"

The Global Water Challenge<\/a> has an initiative called the Water Point Data Exchange<\/a>, which curates and provides data on the locations of these water access points.\u00a0 Currently, there are nearly 500,000 water access points available for 54 countries.\u00a0 Over a quarter of these water points are located in Uganda.<\/p>\n"},{"acf_fc_layout":"image","image":{"ID":1688992,"id":1688992,"title":"Uganda_WaterPoint_map","filename":"Uganda_WaterPoint_map.png","filesize":141912,"url":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/Uganda_WaterPoint_map.png","link":"https:\/\/www.esri.com\/arcgis-blog\/products\/arcgis-pro\/analytics\/automl-in-arcgis-pro-3-0\/uganda_waterpoint_map","alt":"","author":"154341","description":"","caption":"Distribution of water points within Uganda.","name":"uganda_waterpoint_map","status":"inherit","uploaded_to":1688432,"date":"2022-08-18 12:44:04","modified":"2022-08-23 13:27:07","menu_order":0,"mime_type":"image\/png","type":"image","subtype":"png","icon":"https:\/\/www.esri.com\/arcgis-blog\/wp-includes\/images\/media\/default.png","width":519,"height":533,"sizes":{"thumbnail":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/Uganda_WaterPoint_map-213x200.png","thumbnail-width":213,"thumbnail-height":200,"medium":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/Uganda_WaterPoint_map.png","medium-width":254,"medium-height":261,"medium_large":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/Uganda_WaterPoint_map.png","medium_large-width":519,"medium_large-height":533,"large":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/Uganda_WaterPoint_map.png","large-width":519,"large-height":533,"1536x1536":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/Uganda_WaterPoint_map.png","1536x1536-width":519,"1536x1536-height":533,"2048x2048":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/Uganda_WaterPoint_map.png","2048x2048-width":519,"2048x2048-height":533,"card_image":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/Uganda_WaterPoint_map-453x465.png","card_image-width":453,"card_image-height":465,"wide_image":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2022\/08\/Uganda_WaterPoint_map.png","wide_image-width":519,"wide_image-height":533}},"image_position":"center","orientation":"horizontal","hyperlink":""},{"acf_fc_layout":"content","content":"

Using the new AutoML tools in the GeoAI toolbox, I\u2019ll create models that predict whether a water point is functional or not, based on a large number of non-spatial and spatial covariates.\u00a0 The non-spatial covariates are part of the water point dataset, and include:<\/p>\n