One of the nice new features in Django 1.4 is that it allows you to (easily) detect an upstream server's use of ssl using request.is_secure().  This means that your nginx configuration can use http or https on the same server and your Django site can detect it.  The magic is accomplished by *always* setting a header on the nginx side which contains the request protocol on which the request was delivered (in your sites-available/default, for instance):

server {
    listen 80 default deferred;
    listen 443 default ssl;
    server_name myserver.example.com;
    ssl_certificate      /etc/nginx/keys/server.crt;
    ssl_certificate_key  /etc/nginx/keys/server.key;

    location / {
        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
        proxy_set_header Host $http_host;
        proxy_set_header X-Forwarded-Protocol $scheme;
        proxy_redirect off;
        proxy_pass http://app_server;
    }
    ...
}

Then configure Django to consider that header to be an indication of ssl presence in settings.py:

SECURE_PROXY_SSL_HEADER = ('HTTP_X_FORWARDED_PROTOCOL', 'https')

And lastly, generate some (self-signed) ssl certificates for the server (obviously you'll need real certs for a real server):

import socket,subprocess
KEY_FILE = '/etc/nginx/keys/server.key'
CERT_FILE = '/etc/nginx/keys/server.crt'
cn = socket.getfqdn()
# Obviously you should specify your own location...
subject="/C=CA/ST=ON/L=Toronto/CN=%(cn)s"%locals()
subprocess.check_call( [
    'openssl',
        'req',
            '-x509', '-nodes', '-days', '365',
            '-subj', subject,
            '-newkey',
                'rsa:1024',
                '-keyout', KEY_FILE,
                '-out', CERT_FILE,
])

And now request.is_secure() should properly report whether the user is on http or https.  The user will (obviously) get a warning about your self-signed certificate, but better self-signed than nothing, and you can put your real keys on the server when you get them.