import requests
import pandas as pd
import geopandas as gpd
from shapely.geometry import Point
from glob import glob
from geopy.geocoders import NominatimOpenAQ has an aws s3 bucket that contains all the data to download for free. This is a guide to download the data from the bucket. If you have enough space and bandwidth, aws s3 commands are the fastest way to download the data. If you don’t have enough space/bandwidth or you want to download specific data only then follow along.
Acknowledgement: Much help is taken from ChatGPT for some complex Linux commands.
We will mostly use the following commands:
aws s3 ls
aws s3 cpScenario 1: We want to download PurpleAir sensors data for Delhi for entire 2022. I am taking Delhi’s example since there are far lesser sensors in Delhi than in the US. So, it will be easier for this demo.
Some statistics that I have calculated for PurpleAir sensors in US are as follows:
| Country | Number of sensors | Total size | years |
|---|---|---|---|
| USA | 22497 | 90.945 GB | 2018, 2019, 2020, 2021, 2022, 2023 |
Let’s see how to calculate these statistics for India.
!aws s3 --no-sign-request ls s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/ > /tmp/in.txtThe output of above command contains all sensor IDs within India. These sensor IDs are assigned by OpenAQ and they are different from PurpleAir “sensor_index”. The output looks like the following.
with open('/tmp/in.txt', 'r') as f:
lines = [line.strip() for line in f.readlines()]
print(f"Number of locations = {len(lines)}\n")
print(*lines[:3], sep='\n')Number of locations = 623
PRE locationid=160485/
PRE locationid=218334/
PRE locationid=218336/Counting total size of files:
!aws s3 --no-sign-request ls s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/ --recursive > /tmp/in_files.txt!awk '{ sum += $3 } END { print (sum/1024/1024/1024) " GB"}' /tmp/in_files.txt1.20947 GBNumber of years:
!cat /tmp/in_files.txt | grep -oP 'year=\K\w+' | sort | uniq2019
2020
2021
2022
2023Let’s find out which sensors among these belong to Delhi.
!cat /tmp/in.txt | grep -oP 'locationid=\K\w+' | sort | uniq > /tmp/in_locations.txtNow we use OpenAQ REST API. It has limit of 300 requests per 5 minutes. After the limit exceeds, it will return an error.
url = 'https://api.openaq.org/v2/locations'
params = {
'limit': 1000,
'country_id': 'IN',
"modelName": "PurpleAir Sensor"
}
# Request headers
headers = {
'accept': 'application/json'
}
response = requests.get(url, params=params, headers=headers)
if response.status_code == 200:
data = response.json()
else:
print(response.status_code, response.reason)500 Internal Server Errordf = pd.DataFrame(data['results'])
df.shape--------------------------------------------------------------------------- NameError Traceback (most recent call last) Cell In[10], line 1 ----> 1 df = pd.DataFrame(data['results']) 2 df.shape NameError: name 'data' is not defined
The following method works but takes too much time since it is online:
def get_city_name(coords):
latitude = coords["latitude"]
longitude = coords["longitude"]
geolocator = Nominatim(user_agent="myGeocoder") # Replace "myGeocoder" with your desired user agent
location = geolocator.reverse((latitude, longitude), exactly_one=True)
if location:
address = location.raw.get('address', {})
city = address.get('city', '')
return city
return None
df["city"] = df["coordinates"].apply(get_city_name)
df| id | city | name | entity | country | sources | isMobile | isAnalysis | parameters | sensorType | coordinates | lastUpdated | firstUpdated | measurements | bounds | manufacturers | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 318146 | Gangtok | NASA_AQCS_201_cpa | None | IN | None | False | None | [{'id': 135, 'unit': 'particles/cm³', 'count':... | None | {'latitude': 27.31013, 'longitude': 88.59687} | 2023-07-26T02:34:05+00:00 | 2022-04-23T07:42:24+00:00 | 1168071 | [88.59687, 27.31013, 88.59687, 27.31013] | [{'modelName': 'PurpleAir Sensor', 'manufactur... |
| 1 | 220706 | Gangtok | NASA_AQCS_139 | None | IN | None | False | None | [{'id': 100, 'unit': 'c', 'count': 28600, 'ave... | None | {'latitude': 27.310116, 'longitude': 88.59682} | 2023-07-26T02:34:04+00:00 | 2021-02-17T09:56:06+00:00 | 2205110 | [88.59682, 27.310116, 88.59682, 27.310116] | [{'modelName': 'PurpleAir Sensor', 'manufactur... |
| 2 | 66673 | Hisar | NASA_AQCS_160 | None | IN | None | False | None | [{'id': 132, 'unit': 'mb', 'count': 28651, 'av... | None | {'latitude': 29.146254, 'longitude': 75.72236} | 2023-07-26T02:33:56+00:00 | 2021-01-19T23:59:16+00:00 | 2396934 | [75.72236, 29.146254, 75.72236, 29.146254] | [{'modelName': 'PurpleAir Sensor', 'manufactur... |
| 3 | 72977 | Bengaluru | UT Sensor 101 | None | IN | None | False | None | [{'id': 126, 'unit': 'particles/cm³', 'count':... | None | {'latitude': 13.045313, 'longitude': 77.573395} | 2023-07-26T02:33:55+00:00 | 2021-01-14T01:18:23+00:00 | 2498544 | [77.573395, 13.045313, 77.573395, 13.045313] | [{'modelName': 'PurpleAir Sensor', 'manufactur... |
| 4 | 235916 | Bengaluru | UW Sensor 311 | None | IN | None | False | None | [{'id': 130, 'unit': 'particles/cm³', 'count':... | None | {'latitude': 13.048528, 'longitude': 77.582275} | 2023-07-26T02:33:49+00:00 | 2021-09-16T12:29:52+00:00 | 1773924 | [77.582275, 13.048528, 77.582275, 13.048528] | [{'modelName': 'PurpleAir Sensor', 'manufactur... |
| ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... |
| 604 | 73922 | New Delhi District | US Embassy A | None | IN | None | False | None | [{'id': 2, 'unit': 'µg/m³', 'count': 347, 'ave... | None | {'latitude': 28.5979, 'longitude': 77.1847} | 2021-02-05T18:25:34+00:00 | 2021-01-08T12:12:29+00:00 | 2082 | [77.1847, 28.5979, 77.1847, 28.5979] | [{'modelName': 'PurpleAir Sensor', 'manufactur... |
| 605 | 73924 | New Delhi District | US Embassy B | None | IN | None | False | None | [{'id': 2, 'unit': 'µg/m³', 'count': 347, 'ave... | None | {'latitude': 28.5982, 'longitude': 77.1837} | 2021-02-05T18:23:59+00:00 | 2021-01-08T12:11:29+00:00 | 2082 | [77.1837, 28.5982, 77.1837, 28.5982] | [{'modelName': 'PurpleAir Sensor', 'manufactur... |
| 606 | 219370 | Bhubaneswar Municipal Corporation | Bhubaneswar India | None | IN | None | False | None | [{'id': 2, 'unit': 'µg/m³', 'count': 180, 'ave... | None | {'latitude': 20.2853, 'longitude': 85.7685} | 2021-02-03T10:18:41+00:00 | 2021-02-03T03:02:39+00:00 | 1080 | [85.7685, 20.2853, 85.7685, 20.2853] | [{'modelName': 'PurpleAir Sensor', 'manufactur... |
| 607 | 71465 | Gurugram District | NASA_AQCS_152 | None | IN | None | False | None | [{'id': 1, 'unit': 'µg/m³', 'count': 1, 'avera... | None | {'latitude': 28.4522, 'longitude': 77.0949} | 2021-01-14T01:18:54+00:00 | 2021-01-14T01:18:54+00:00 | 6 | [77.0949, 28.4522, 77.0949, 28.4522] | [{'modelName': 'PurpleAir Sensor', 'manufactur... |
| 608 | 221974 | Bengaluru | UT sensor | None | IN | None | False | None | [{'id': 130, 'unit': 'particles/cm³', 'count':... | None | {'latitude': 13.0449, 'longitude': 77.5788} | 2019-12-17T10:32:35+00:00 | 2019-12-17T10:32:35+00:00 | 6 | [77.5788, 13.0449, 77.5788, 13.0449] | [{'modelName': 'PurpleAir Sensor', 'manufactur... |
609 rows × 16 columns
Now, we use another method of shapefile to do this:
!wget --no-check-certificate "https://groups.google.com/group/datameet/attach/29b74b1aef5f2f13/Delhi.zip?part=0.1" -O /tmp/delhi.zip
!unzip -o /tmp/delhi.zip -d /tmp/delhi--2023-07-26 08:37:34-- https://groups.google.com/group/datameet/attach/29b74b1aef5f2f13/Delhi.zip?part=0.1
Resolving groups.google.com (groups.google.com)... 216.239.32.177, 216.239.36.177, 216.239.38.177, ...
Connecting to groups.google.com (groups.google.com)|216.239.32.177|:443... connected.
HTTP request sent, awaiting response... 302 Moved Temporarily
Location: https://06895207363394598426.googlegroups.com/attach/29b74b1aef5f2f13/Delhi.zip?part=0.1&vt=ANaJVrEpUPnptnb4Y-J5gJRBVJ29K0pIGKzeBG7492Ume1tyn1MY5eTDbztxP0Hdbc7u8XhmH_GbemY_HD60x5OvDhr7M2ib1h8YfDmlNxFefazGPgmAUj0 [following]
--2023-07-26 08:37:35-- https://06895207363394598426.googlegroups.com/attach/29b74b1aef5f2f13/Delhi.zip?part=0.1&vt=ANaJVrEpUPnptnb4Y-J5gJRBVJ29K0pIGKzeBG7492Ume1tyn1MY5eTDbztxP0Hdbc7u8XhmH_GbemY_HD60x5OvDhr7M2ib1h8YfDmlNxFefazGPgmAUj0
Resolving 06895207363394598426.googlegroups.com (06895207363394598426.googlegroups.com)... 142.251.10.137, 2404:6800:4003:c0f::89
Connecting to 06895207363394598426.googlegroups.com (06895207363394598426.googlegroups.com)|142.251.10.137|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: unspecified [application/zip]
Saving to: ‘/tmp/delhi.zip’
/tmp/delhi.zip [ <=> ] 16.42K --.-KB/s in 0.04s
2023-07-26 08:37:37 (429 KB/s) - ‘/tmp/delhi.zip’ saved [16812]
Archive: /tmp/delhi.zip
inflating: /tmp/delhi/Delhi.kml
inflating: /tmp/delhi/Districts.dbf
inflating: /tmp/delhi/Districts.prj
inflating: /tmp/delhi/Districts.qpj
inflating: /tmp/delhi/Districts.shp
inflating: /tmp/delhi/Districts.shx gdf = gpd.read_file('/tmp/delhi/Districts.shp')
gdf.plot(color="none", edgecolor="black");
# check if a point is within Delhi
def is_within_delhi(coords):
point = Point(coords["longitude"], coords["latitude"])
for i, row in gdf.iterrows():
if row.geometry.contains(point):
return True
return False
df["is_within_delhi"] = df["coordinates"].apply(is_within_delhi)delhi_df = df[df["is_within_delhi"]]
delhi_df.shape(311, 17)delhi_df.city.value_counts() 194
New Delhi District 112
Dwarka 3
Ghaziabad 2
Name: city, dtype: int64Seems like many points were not detected by the online geopy encoder.
Now, we know that out of 623, 311 sensors belong to Delhi. Let’s download the data for these sensors. For illustration, I will download data for 3 sensors for year 2022 and month of Jan.
# dump delhi_df.id to a file
delhi_df.id.to_csv('/tmp/delhi_locations.txt', index=False, header=False)
!head -n3 /tmp/delhi_locations.txt274208
221227
273205!head -n3 /tmp/delhi_locations.txt > /tmp/delhi_locations_3.txt
!while read -r sensor_id; do aws s3 --no-sign-request cp s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=$sensor_id/year=2022/month=01 /tmp/delhi_data --recursive; done < /tmp/delhi_locations_3.txtdownload: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=274208/year=2022/month=01/location-274208-20220107.csv.gz to ../../../../tmp/delhi_data/location-274208-20220107.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=274208/year=2022/month=01/location-274208-20220120.csv.gz to ../../../../tmp/delhi_data/location-274208-20220120.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=274208/year=2022/month=01/location-274208-20220131.csv.gz to ../../../../tmp/delhi_data/location-274208-20220131.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=274208/year=2022/month=01/location-274208-20220122.csv.gz to ../../../../tmp/delhi_data/location-274208-20220122.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=274208/year=2022/month=01/location-274208-20220130.csv.gz to ../../../../tmp/delhi_data/location-274208-20220130.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=274208/year=2022/month=01/location-274208-20220129.csv.gz to ../../../../tmp/delhi_data/location-274208-20220129.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=274208/year=2022/month=01/location-274208-20220114.csv.gz to ../../../../tmp/delhi_data/location-274208-20220114.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=274208/year=2022/month=01/location-274208-20220123.csv.gz to ../../../../tmp/delhi_data/location-274208-20220123.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=274208/year=2022/month=01/location-274208-20220113.csv.gz to ../../../../tmp/delhi_data/location-274208-20220113.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=274208/year=2022/month=01/location-274208-20220121.csv.gz to ../../../../tmp/delhi_data/location-274208-20220121.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=221227/year=2022/month=01/location-221227-20220107.csv.gz to ../../../../tmp/delhi_data/location-221227-20220107.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=221227/year=2022/month=01/location-221227-20220104.csv.gz to ../../../../tmp/delhi_data/location-221227-20220104.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=221227/year=2022/month=01/location-221227-20220115.csv.gz to ../../../../tmp/delhi_data/location-221227-20220115.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=221227/year=2022/month=01/location-221227-20220105.csv.gz to ../../../../tmp/delhi_data/location-221227-20220105.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=221227/year=2022/month=01/location-221227-20220108.csv.gz to ../../../../tmp/delhi_data/location-221227-20220108.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=221227/year=2022/month=01/location-221227-20220106.csv.gz to ../../../../tmp/delhi_data/location-221227-20220106.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=221227/year=2022/month=01/location-221227-20220102.csv.gz to ../../../../tmp/delhi_data/location-221227-20220102.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=221227/year=2022/month=01/location-221227-20220117.csv.gz to ../../../../tmp/delhi_data/location-221227-20220117.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=221227/year=2022/month=01/location-221227-20220118.csv.gz to ../../../../tmp/delhi_data/location-221227-20220118.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=221227/year=2022/month=01/location-221227-20220110.csv.gz to ../../../../tmp/delhi_data/location-221227-20220110.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=221227/year=2022/month=01/location-221227-20220131.csv.gz to ../../../../tmp/delhi_data/location-221227-20220131.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=221227/year=2022/month=01/location-221227-20220113.csv.gz to ../../../../tmp/delhi_data/location-221227-20220113.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=221227/year=2022/month=01/location-221227-20220119.csv.gz to ../../../../tmp/delhi_data/location-221227-20220119.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=221227/year=2022/month=01/location-221227-20220121.csv.gz to ../../../../tmp/delhi_data/location-221227-20220121.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=221227/year=2022/month=01/location-221227-20220123.csv.gz to ../../../../tmp/delhi_data/location-221227-20220123.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=221227/year=2022/month=01/location-221227-20220116.csv.gz to ../../../../tmp/delhi_data/location-221227-20220116.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=221227/year=2022/month=01/location-221227-20220128.csv.gz to ../../../../tmp/delhi_data/location-221227-20220128.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=221227/year=2022/month=01/location-221227-20220122.csv.gz to ../../../../tmp/delhi_data/location-221227-20220122.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=221227/year=2022/month=01/location-221227-20220101.csv.gz to ../../../../tmp/delhi_data/location-221227-20220101.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=221227/year=2022/month=01/location-221227-20220111.csv.gz to ../../../../tmp/delhi_data/location-221227-20220111.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=221227/year=2022/month=01/location-221227-20220120.csv.gz to ../../../../tmp/delhi_data/location-221227-20220120.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=221227/year=2022/month=01/location-221227-20220109.csv.gz to ../../../../tmp/delhi_data/location-221227-20220109.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=221227/year=2022/month=01/location-221227-20220129.csv.gz to ../../../../tmp/delhi_data/location-221227-20220129.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=221227/year=2022/month=01/location-221227-20220103.csv.gz to ../../../../tmp/delhi_data/location-221227-20220103.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=221227/year=2022/month=01/location-221227-20220112.csv.gz to ../../../../tmp/delhi_data/location-221227-20220112.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=221227/year=2022/month=01/location-221227-20220114.csv.gz to ../../../../tmp/delhi_data/location-221227-20220114.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=221227/year=2022/month=01/location-221227-20220130.csv.gz to ../../../../tmp/delhi_data/location-221227-20220130.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=273205/year=2022/month=01/location-273205-20220120.csv.gz to ../../../../tmp/delhi_data/location-273205-20220120.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=273205/year=2022/month=01/location-273205-20220124.csv.gz to ../../../../tmp/delhi_data/location-273205-20220124.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=273205/year=2022/month=01/location-273205-20220107.csv.gz to ../../../../tmp/delhi_data/location-273205-20220107.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=273205/year=2022/month=01/location-273205-20220121.csv.gz to ../../../../tmp/delhi_data/location-273205-20220121.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=273205/year=2022/month=01/location-273205-20220126.csv.gz to ../../../../tmp/delhi_data/location-273205-20220126.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=273205/year=2022/month=01/location-273205-20220114.csv.gz to ../../../../tmp/delhi_data/location-273205-20220114.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=273205/year=2022/month=01/location-273205-20220129.csv.gz to ../../../../tmp/delhi_data/location-273205-20220129.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=273205/year=2022/month=01/location-273205-20220130.csv.gz to ../../../../tmp/delhi_data/location-273205-20220130.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=273205/year=2022/month=01/location-273205-20220113.csv.gz to ../../../../tmp/delhi_data/location-273205-20220113.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=273205/year=2022/month=01/location-273205-20220123.csv.gz to ../../../../tmp/delhi_data/location-273205-20220123.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=273205/year=2022/month=01/location-273205-20220122.csv.gz to ../../../../tmp/delhi_data/location-273205-20220122.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=273205/year=2022/month=01/location-273205-20220131.csv.gz to ../../../../tmp/delhi_data/location-273205-20220131.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=273205/year=2022/month=01/location-273205-20220128.csv.gz to ../../../../tmp/delhi_data/location-273205-20220128.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=273205/year=2022/month=01/location-273205-20220125.csv.gz to ../../../../tmp/delhi_data/location-273205-20220125.csv.gz
download: s3://openaq-data-archive/records/csv.gz/provider=purpleair/country=in/locationid=273205/year=2022/month=01/location-273205-20220127.csv.gz to ../../../../tmp/delhi_data/location-273205-20220127.csv.gzVerify if we got all the sensors data as we needed.
!ls /tmp/delhi_data/location-221227* | wc -l
!ls /tmp/delhi_data/location-274208* | wc -l
!ls /tmp/delhi_data/location-273205* | wc -l27
10
15sensor_df = pd.read_csv('/tmp/delhi_data/location-274208-20220107.csv.gz')
sensor_df.parameter.value_counts()pm10 64
pm25 64
pm1 64
um010 64
um025 64
um100 64
Name: parameter, dtype: int64