jupyter:

jupytext:

notebook_metadata_filter: all

text_representation:

extension: .md

format_name: markdown

format_version: '1.1'

jupytext_version: 1.1.7

kernelspec:

display_name: Python 3

language: python

name: python3

language_info:

codemirror_mode:

name: ipython

version: 3

file_extension: .py

mimetype: text/x-python

name: python

nbconvert_exporter: python

pygments_lexer: ipython3

version: 3.6.5

plotly:

permalink: python/data-api/

description: How to upload data to Plotly from Python with the Plotly Grid API.

name: Upload Data to Plotly from Python

has_thumbnail: true

thumbnail: thumbnail/table.jpg

layout: user-guide

name: Plots from Grids

language: python

title: Plotly Data API

display_as: chart_studio

has_thumbnail: true

page_type: u-guide

order: 5

v4upgrade: true

You can instantiate a grid with data by either uploading tabular data to Plotly or by creating a Plotly `grid` using the API. To upload the grid we will use `plotly.plotly.grid_ops.upload()`. It takes the following arguments:

- `grid` (Grid Object): the actual grid object that you are uploading.

- `filename` (str): name of the grid in your plotly account,

- `world_readable` (bool): if `True`, the grid is `public` and can be viewed by anyone in your files. If `False`, it is private and can only be viewed by you.

- `auto_open` (bool): if determines if the grid is opened in the browser or not.

You can run `help(py.grid_ops.upload)` for a more detailed description of these and all the arguments.

import chart_studio

import chart_studio.plotly as py

import chart_studio.tools as tls

import plotly.graph_objects as go

from chart_studio.grid_objs import Column, Grid

from datetime import datetime as dt

import numpy as np

from IPython.display import IFrame

column_1 = Column(['a', 'b', 'c'], 'column 1')

column_2 = Column([1, 2, 3], 'column 2') # Tabular data can be numbers, strings, or dates

grid = Grid([column_1, column_2])

url = py.grid_ops.upload(grid,

filename='grid_ex_'+str(dt.now()),

world_readable=True,

auto_open=False)

print(url)

You can view your newly created grid at the `url`:

IFrame(src= url.rstrip('/') + ".embed", width="100%",height="200px", frameBorder="0")

You are also able to view the grid in your list of files inside your [organize folder](https://plot.ly/organize).

Along with uploading a grid, you can upload a Dataframe as well as convert it to raw data as a grid:

import chart_studio.plotly as py

import plotly.figure_factory as ff

import pandas as pd

df = pd.read_csv('https://raw.githubusercontent.com/plotly/datasets/master/2014_apple_stock.csv')

df_head = df.head()

table = ff.create_table(df_head)

py.iplot(table, filename='dataframe_ex_preview')

Plotly graphs are usually described with data embedded in them. For example, here we place `x` and `y` data directly into our `Histogram2dContour` object:

x = np.random.randn(1000)

y = np.random.randn(1000) + 1

data = [

go.Histogram2dContour(

x=x,

y=y

)

]

py.iplot(data, filename='Example 2D Histogram Contour')

We can also create graphs based off of references to columns of grids. Here, we'll upload several `column`s to our Plotly account:

column_1 = Column(np.random.randn(1000), 'column 1')

column_2 = Column(np.random.randn(1000)+1, 'column 2')

column_3 = Column(np.random.randn(1000)+2, 'column 3')

column_4 = Column(np.random.randn(1000)+3, 'column 4')

grid = Grid([column_1, column_2, column_3, column_4])

url = py.grid_ops.upload(grid, filename='randn_int_offset_'+str(dt.now()))

print(url)

IFrame(src= url.rstrip('/') + ".embed", width="100%",height="200px", frameBorder="0")

Instead of placing data into `x` and `y`, we'll place our Grid columns into `xsrc` and `ysrc`:

data = [

go.Histogram2dContour(

xsrc=grid[0],

ysrc=grid[1]

)

]

py.iplot(data, filename='2D Contour from Grid Data')

So, when you view the data, you'll see your original grid, not just the columns that compose this graph:

In [Chart Studio Enterprise](https://plot.ly/product/enterprise/), you can upload and assign free-form JSON `metadata` to any grid object. This means that you can keep all of your raw data in one place, under one grid.

If you update the original data source, in the workspace or with our API, all of the graphs that are sourced from it will be updated as well. You can make multiple graphs from a single Grid and you can make a graph from multiple grids. You can also add rows and columns to existing grids programatically.

meta = {

"Month": "November",

"Experiment ID": "d3kbd",

"Operator": "James Murphy",

"Initial Conditions": {

"Voltage": 5.5

}

}

grid_url = py.grid_ops.upload(grid, filename='grid_with_metadata_'+str(dt.now()), meta=meta)

print(url)

help(py.grid_ops)